<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/">
  <channel>
    <title>Kreath Archive</title>
    <link>https://archive.kreathlab.com</link>
    <description>나를 기록하고 보여주는 공간 — 기술 블로그, 포트폴리오, 독서 아카이브</description>
    <language>ko</language>
    <lastBuildDate>Wed, 08 Apr 2026 13:12:23 GMT</lastBuildDate>
    <atom:link href="https://archive.kreathlab.com/feed.xml" rel="self" type="application/rss+xml"/>
    <managingEditor>hello@kreathlab.com (Kreath)</managingEditor>
    <item>
      <title><![CDATA[10장: 실전 프로젝트 — 스트리밍 AI 시스템 구축]]></title>
      <link>https://archive.kreathlab.com/tech/streaming-ai-10-project</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/streaming-ai-10-project</guid>
      <description><![CDATA[SSE, gRPC, WebSocket을 결합한 하이브리드 스트리밍 AI 시스템을 설계하고 구현합니다. 프로토콜 선택 의사결정 트리, 엔드투엔드 구현, 성능 최적화, 운영 체크리스트를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>SSE, gRPC, WebSocket을 결합한 하이브리드 스트리밍 AI 시스템을 설계하고 구현합니다. 프로토콜 선택 의사결정 트리, 엔드투엔드 구현, 성능 최적화, 운영 체크리스트를 다룹니다.</p><p><strong>아키텍처</strong> / 21분 읽기 / 시리즈: streaming-ai</p><p><a href="https://archive.kreathlab.com/tech/streaming-ai-10-project">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 05 Apr 2026 00:00:00 GMT</pubDate>
      <category>architecture</category>
      <category>streaming</category>
      <category>ai</category>
    </item>
    <item>
      <title><![CDATA[9장: 프로덕션 AI 데이터 파이프라인]]></title>
      <link>https://archive.kreathlab.com/tech/structured-output-09-production</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/structured-output-09-production</guid>
      <description><![CDATA[재시도, 서킷 브레이커, 관측 가능성, 비용 추적, 스키마 버전 관리 등 프로덕션 수준의 AI 파이프라인 운영 기법을 학습합니다.]]></description>
      <content:encoded><![CDATA[<p>재시도, 서킷 브레이커, 관측 가능성, 비용 추적, 스키마 버전 관리 등 프로덕션 수준의 AI 파이프라인 운영 기법을 학습합니다.</p><p><strong>AI / ML</strong> / 15분 읽기 / 시리즈: structured-output</p><p><a href="https://archive.kreathlab.com/tech/structured-output-09-production">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 05 Apr 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>structured-output</category>
      <category>ai</category>
      <category>data-engineering</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[10장: 실전 프로젝트 — Structured Output 파이프라인 구축]]></title>
      <link>https://archive.kreathlab.com/tech/structured-output-10-project</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/structured-output-10-project</guid>
      <description><![CDATA[PDF 송장에서 구조화된 JSON 데이터를 추출하는 엔드투엔드 파이프라인을 FastAPI, Pydantic, 검증 루프, 배치 처리로 구축합니다.]]></description>
      <content:encoded><![CDATA[<p>PDF 송장에서 구조화된 JSON 데이터를 추출하는 엔드투엔드 파이프라인을 FastAPI, Pydantic, 검증 루프, 배치 처리로 구축합니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: structured-output</p><p><a href="https://archive.kreathlab.com/tech/structured-output-10-project">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 05 Apr 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>structured-output</category>
      <category>ai</category>
      <category>data-engineering</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[9장: 합성 데이터 평가와 벤치마킹]]></title>
      <link>https://archive.kreathlab.com/tech/synthetic-data-09-evaluation</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/synthetic-data-09-evaluation</guid>
      <description><![CDATA[TSTR 방법론, 다양성 메트릭, 분포 비교, 다운스트림 성능 측정, 합성 vs 실제 데이터 비교 실험, 벤치마크 설계 방법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>TSTR 방법론, 다양성 메트릭, 분포 비교, 다운스트림 성능 측정, 합성 vs 실제 데이터 비교 실험, 벤치마크 설계 방법을 다룹니다.</p><p><strong>AI / ML</strong> / 21분 읽기 / 시리즈: synthetic-data</p><p><a href="https://archive.kreathlab.com/tech/synthetic-data-09-evaluation">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 05 Apr 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>synthetic-data</category>
      <category>ai</category>
      <category>data-engineering</category>
      <category>llm</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[10장: 실전 프로젝트 -- 합성 데이터 파이프라인 구축]]></title>
      <link>https://archive.kreathlab.com/tech/synthetic-data-10-pipeline</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/synthetic-data-10-pipeline</guid>
      <description><![CDATA[엔드투엔드 합성 데이터 파이프라인 아키텍처, 생성-검증-필터링-증강-평가 통합, CI/CD 연동, 자동화된 품질 게이트, 비용 최적화, 프로덕션 운영 전략을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>엔드투엔드 합성 데이터 파이프라인 아키텍처, 생성-검증-필터링-증강-평가 통합, CI/CD 연동, 자동화된 품질 게이트, 비용 최적화, 프로덕션 운영 전략을 다룹니다.</p><p><strong>AI / ML</strong> / 24분 읽기 / 시리즈: synthetic-data</p><p><a href="https://archive.kreathlab.com/tech/synthetic-data-10-pipeline">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 05 Apr 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>synthetic-data</category>
      <category>ai</category>
      <category>data-engineering</category>
      <category>llm</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[8장: 서버사이드 Wasm — Spin, Fermyon, SpinKube]]></title>
      <link>https://archive.kreathlab.com/tech/webassembly-08-serverside</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/webassembly-08-serverside</guid>
      <description><![CDATA[Spin 프레임워크의 아키텍처, Fermyon Cloud와 Akamai 통합, SpinKube를 활용한 Kubernetes 배포, Docker와 Wasm의 비교를 통해 서버사이드 WebAssembly의 현재를 분석합니다.]]></description>
      <content:encoded><![CDATA[<p>Spin 프레임워크의 아키텍처, Fermyon Cloud와 Akamai 통합, SpinKube를 활용한 Kubernetes 배포, Docker와 Wasm의 비교를 통해 서버사이드 WebAssembly의 현재를 분석합니다.</p><p><strong>프로그래밍</strong> / 16분 읽기 / 시리즈: webassembly</p><p><a href="https://archive.kreathlab.com/tech/webassembly-08-serverside">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 05 Apr 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>webassembly</category>
      <category>rust</category>
    </item>
    <item>
      <title><![CDATA[9장: 엣지 컴퓨팅과 Wasm 배포]]></title>
      <link>https://archive.kreathlab.com/tech/webassembly-09-edge</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/webassembly-09-edge</guid>
      <description><![CDATA[Cloudflare Workers, Fastly Compute, Akamai Edge의 Wasm 실행 환경을 비교하고, 엣지에서의 AI 추론, 콜드 스타트 성능 분석, 엣지 배포 전략을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Cloudflare Workers, Fastly Compute, Akamai Edge의 Wasm 실행 환경을 비교하고, 엣지에서의 AI 추론, 콜드 스타트 성능 분석, 엣지 배포 전략을 다룹니다.</p><p><strong>프로그래밍</strong> / 16분 읽기 / 시리즈: webassembly</p><p><a href="https://archive.kreathlab.com/tech/webassembly-09-edge">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 05 Apr 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>webassembly</category>
      <category>rust</category>
    </item>
    <item>
      <title><![CDATA[10장: 실전 프로젝트 — WebAssembly 애플리케이션 구축]]></title>
      <link>https://archive.kreathlab.com/tech/webassembly-10-project</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/webassembly-10-project</guid>
      <description><![CDATA[Rust+Spin 서버리스 API 구축, 브라우저 Wasm 모듈 통합, 엣지 배포, 성능 벤치마킹, Wasm 도입 의사결정 가이드, 그리고 WebAssembly의 미래 전망을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Rust+Spin 서버리스 API 구축, 브라우저 Wasm 모듈 통합, 엣지 배포, 성능 벤치마킹, Wasm 도입 의사결정 가이드, 그리고 WebAssembly의 미래 전망을 다룹니다.</p><p><strong>프로그래밍</strong> / 19분 읽기 / 시리즈: webassembly</p><p><a href="https://archive.kreathlab.com/tech/webassembly-10-project">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 05 Apr 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>webassembly</category>
      <category>rust</category>
    </item>
    <item>
      <title><![CDATA[10장: 실전 프로젝트 -- Platform Engineering 구축]]></title>
      <link>https://archive.kreathlab.com/tech/platform-engineering-10-project</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/platform-engineering-10-project</guid>
      <description><![CDATA[Backstage, ArgoCD, Crossplane으로 엔드투엔드 IDP를 구축하는 실전 프로젝트. Golden Path 작성, 셀프서비스 워크플로우, 비용 가시성 통합까지 전 과정을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Backstage, ArgoCD, Crossplane으로 엔드투엔드 IDP를 구축하는 실전 프로젝트. Golden Path 작성, 셀프서비스 워크플로우, 비용 가시성 통합까지 전 과정을 다룹니다.</p><p><strong>인프라</strong> / 17분 읽기 / 시리즈: platform-engineering</p><p><a href="https://archive.kreathlab.com/tech/platform-engineering-10-project">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 04 Apr 2026 00:00:00 GMT</pubDate>
      <category>infra</category>
      <category>platform-engineering</category>
      <category>devops</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[8장: 도메인 특화 데이터셋 구축]]></title>
      <link>https://archive.kreathlab.com/tech/synthetic-data-08-domain-specific</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/synthetic-data-08-domain-specific</guid>
      <description><![CDATA[의료, 법률, 금융, 코드 도메인별 합성 데이터 접근법, 전문가 시드 데이터 설계, InstructLab 택소노미 방식, 도메인 검증 전략을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>의료, 법률, 금융, 코드 도메인별 합성 데이터 접근법, 전문가 시드 데이터 설계, InstructLab 택소노미 방식, 도메인 검증 전략을 다룹니다.</p><p><strong>AI / ML</strong> / 22분 읽기 / 시리즈: synthetic-data</p><p><a href="https://archive.kreathlab.com/tech/synthetic-data-08-domain-specific">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 04 Apr 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>synthetic-data</category>
      <category>ai</category>
      <category>data-engineering</category>
      <category>llm</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[10장: 프로덕션 추론 최적화 실전]]></title>
      <link>https://archive.kreathlab.com/tech/llm-inference-10-production</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/llm-inference-10-production</guid>
      <description><![CDATA[FP8, FlashAttention 3, Continuous Batching, Speculative Decoding을 조합해 5-8배 비용 효율을 달성하는 실전 배포 파이프라인과 모니터링 전략을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>FP8, FlashAttention 3, Continuous Batching, Speculative Decoding을 조합해 5-8배 비용 효율을 달성하는 실전 배포 파이프라인과 모니터링 전략을 다룹니다.</p><p><strong>AI / ML</strong> / 20분 읽기 / 시리즈: llm-inference</p><p><a href="https://archive.kreathlab.com/tech/llm-inference-10-production">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 03 Apr 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>ai</category>
      <category>performance</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[9장: 프로덕션 스트리밍 인프라]]></title>
      <link>https://archive.kreathlab.com/tech/streaming-ai-09-infrastructure</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/streaming-ai-09-infrastructure</guid>
      <description><![CDATA[로드밸런서의 WebSocket 업그레이드, CDN과 스트리밍, Kubernetes에서의 스트리밍 서비스 운영, 모니터링 전략, HTTP/3(QUIC)과 WebTransport의 미래를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>로드밸런서의 WebSocket 업그레이드, CDN과 스트리밍, Kubernetes에서의 스트리밍 서비스 운영, 모니터링 전략, HTTP/3(QUIC)과 WebTransport의 미래를 다룹니다.</p><p><strong>아키텍처</strong> / 17분 읽기 / 시리즈: streaming-ai</p><p><a href="https://archive.kreathlab.com/tech/streaming-ai-09-infrastructure">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 03 Apr 2026 00:00:00 GMT</pubDate>
      <category>architecture</category>
      <category>streaming</category>
      <category>ai</category>
    </item>
    <item>
      <title><![CDATA[8장: 출력 검증과 폴백 전략]]></title>
      <link>https://archive.kreathlab.com/tech/structured-output-08-validation</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/structured-output-08-validation</guid>
      <description><![CDATA[스키마 검증, 의미적 검증, 자동 재시도, 멀티 프로바이더 폴백, 부분 출력 복구 등 프로덕션 수준의 검증 전략을 학습합니다.]]></description>
      <content:encoded><![CDATA[<p>스키마 검증, 의미적 검증, 자동 재시도, 멀티 프로바이더 폴백, 부분 출력 복구 등 프로덕션 수준의 검증 전략을 학습합니다.</p><p><strong>AI / ML</strong> / 18분 읽기 / 시리즈: structured-output</p><p><a href="https://archive.kreathlab.com/tech/structured-output-08-validation">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 03 Apr 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>structured-output</category>
      <category>ai</category>
      <category>data-engineering</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[7장: 브라우저 고성능 앱 — Wasm의 원래 영역]]></title>
      <link>https://archive.kreathlab.com/tech/webassembly-07-browser</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/webassembly-07-browser</guid>
      <description><![CDATA[JavaScript와 Wasm의 상호 호출, Web API 연동, 이미지/비디오 처리, AI 추론, 게임 엔진 등 브라우저에서 WebAssembly로 고성능 애플리케이션을 구축하는 방법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>JavaScript와 Wasm의 상호 호출, Web API 연동, 이미지/비디오 처리, AI 추론, 게임 엔진 등 브라우저에서 WebAssembly로 고성능 애플리케이션을 구축하는 방법을 다룹니다.</p><p><strong>프로그래밍</strong> / 14분 읽기 / 시리즈: webassembly</p><p><a href="https://archive.kreathlab.com/tech/webassembly-07-browser">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 03 Apr 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>webassembly</category>
      <category>rust</category>
    </item>
    <item>
      <title><![CDATA[10장: 실전 프로젝트 — Knowledge Graph + AI 시스템]]></title>
      <link>https://archive.kreathlab.com/tech/knowledge-graph-10-project</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/knowledge-graph-10-project</guid>
      <description><![CDATA[기술 문서에서 LLM으로 지식 그래프를 구축하고, GraphRAG로 자연어 질의를 처리하며, 벡터 전용 RAG와 성능을 비교하는 엔드투엔드 실전 프로젝트를 구현합니다.]]></description>
      <content:encoded><![CDATA[<p>기술 문서에서 LLM으로 지식 그래프를 구축하고, GraphRAG로 자연어 질의를 처리하며, 벡터 전용 RAG와 성능을 비교하는 엔드투엔드 실전 프로젝트를 구현합니다.</p><p><strong>AI / ML</strong> / 23분 읽기 / 시리즈: knowledge-graph</p><p><a href="https://archive.kreathlab.com/tech/knowledge-graph-10-project">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 02 Apr 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>knowledge-graph</category>
      <category>ai</category>
      <category>data-engineering</category>
    </item>
    <item>
      <title><![CDATA[9장: 조직 확장과 플랫폼 팀 운영]]></title>
      <link>https://archive.kreathlab.com/tech/platform-engineering-09-scaling</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/platform-engineering-09-scaling</guid>
      <description><![CDATA[플랫폼 팀의 구조와 Team Topologies 적용, 채택률 측정과 개선, 개발자 만족도(NPS), 이해관계자 관리, 그리고 플랫폼 성숙도 모델을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>플랫폼 팀의 구조와 Team Topologies 적용, 채택률 측정과 개선, 개발자 만족도(NPS), 이해관계자 관리, 그리고 플랫폼 성숙도 모델을 다룹니다.</p><p><strong>인프라</strong> / 22분 읽기 / 시리즈: platform-engineering</p><p><a href="https://archive.kreathlab.com/tech/platform-engineering-09-scaling">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 02 Apr 2026 00:00:00 GMT</pubDate>
      <category>infra</category>
      <category>platform-engineering</category>
      <category>devops</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[7장: 프라이버시 보존 합성 데이터]]></title>
      <link>https://archive.kreathlab.com/tech/synthetic-data-07-privacy</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/synthetic-data-07-privacy</guid>
      <description><![CDATA[차등 프라이버시, PII 마스킹, 멤버십 추론 공격 방어, 유사도 필터, 규제 대응 전략과 프라이버시-유용성 트레이드오프를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>차등 프라이버시, PII 마스킹, 멤버십 추론 공격 방어, 유사도 필터, 규제 대응 전략과 프라이버시-유용성 트레이드오프를 다룹니다.</p><p><strong>AI / ML</strong> / 20분 읽기 / 시리즈: synthetic-data</p><p><a href="https://archive.kreathlab.com/tech/synthetic-data-07-privacy">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 02 Apr 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>synthetic-data</category>
      <category>ai</category>
      <category>data-engineering</category>
      <category>llm</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[9장: 비용 vs 지연시간 트레이드오프]]></title>
      <link>https://archive.kreathlab.com/tech/llm-inference-09-cost-latency</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/llm-inference-09-cost-latency</guid>
      <description><![CDATA[GPU 선택, 클라우드 vs 온프레미스, 배칭 전략이 비용과 지연시간에 미치는 영향을 분석하고, SLO 기반 최적화와 비용 모델링 방법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>GPU 선택, 클라우드 vs 온프레미스, 배칭 전략이 비용과 지연시간에 미치는 영향을 분석하고, SLO 기반 최적화와 비용 모델링 방법을 다룹니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: llm-inference</p><p><a href="https://archive.kreathlab.com/tech/llm-inference-09-cost-latency">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 01 Apr 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>ai</category>
      <category>performance</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[8장: 백프레셔와 흐름 제어]]></title>
      <link>https://archive.kreathlab.com/tech/streaming-ai-08-backpressure</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/streaming-ai-08-backpressure</guid>
      <description><![CDATA[생산자-소비자 속도 불일치를 관리하는 백프레셔의 원리, 버퍼링/드롭/속도 제한 전략, LLM API 레이트 리미팅, 토큰 버킷 알고리즘, 큐 깊이 모니터링을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>생산자-소비자 속도 불일치를 관리하는 백프레셔의 원리, 버퍼링/드롭/속도 제한 전략, LLM API 레이트 리미팅, 토큰 버킷 알고리즘, 큐 깊이 모니터링을 다룹니다.</p><p><strong>아키텍처</strong> / 18분 읽기 / 시리즈: streaming-ai</p><p><a href="https://archive.kreathlab.com/tech/streaming-ai-08-backpressure">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 01 Apr 2026 00:00:00 GMT</pubDate>
      <category>architecture</category>
      <category>streaming</category>
      <category>ai</category>
    </item>
    <item>
      <title><![CDATA[7장: ETL 파이프라인에 LLM 통합]]></title>
      <link>https://archive.kreathlab.com/tech/structured-output-07-etl</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/structured-output-07-etl</guid>
      <description><![CDATA[전통 ETL과 LLM-enhanced ETL을 비교하고, Transform 단계에 LLM을 적용하여 분류, 요약, 정규화, 감성분석을 수행하는 방법을 학습합니다.]]></description>
      <content:encoded><![CDATA[<p>전통 ETL과 LLM-enhanced ETL을 비교하고, Transform 단계에 LLM을 적용하여 분류, 요약, 정규화, 감성분석을 수행하는 방법을 학습합니다.</p><p><strong>AI / ML</strong> / 14분 읽기 / 시리즈: structured-output</p><p><a href="https://archive.kreathlab.com/tech/structured-output-07-etl">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 01 Apr 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>structured-output</category>
      <category>ai</category>
      <category>data-engineering</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[6장: Go, Python, 기타 언어에서 Wasm]]></title>
      <link>https://archive.kreathlab.com/tech/webassembly-06-other-languages</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/webassembly-06-other-languages</guid>
      <description><![CDATA[TinyGo, Python(componentize-py), C/C++(Emscripten), AssemblyScript, .NET Blazor 등 다양한 언어의 Wasm 지원 현황과 제약 사항, 언어 선택 가이드를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>TinyGo, Python(componentize-py), C/C++(Emscripten), AssemblyScript, .NET Blazor 등 다양한 언어의 Wasm 지원 현황과 제약 사항, 언어 선택 가이드를 다룹니다.</p><p><strong>프로그래밍</strong> / 14분 읽기 / 시리즈: webassembly</p><p><a href="https://archive.kreathlab.com/tech/webassembly-06-other-languages">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 01 Apr 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>webassembly</category>
      <category>rust</category>
    </item>
    <item>
      <title><![CDATA[9장: 프로덕션 파이프라인 구축]]></title>
      <link>https://archive.kreathlab.com/tech/knowledge-graph-09-production</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/knowledge-graph-09-production</guid>
      <description><![CDATA[지식 그래프의 증분 업데이트, 데이터 품질 검증, 스케일링 전략, 모니터링, 비용 최적화, 그리고 Graphiti를 활용한 실시간 KG 업데이트까지 프로덕션 운영의 핵심을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>지식 그래프의 증분 업데이트, 데이터 품질 검증, 스케일링 전략, 모니터링, 비용 최적화, 그리고 Graphiti를 활용한 실시간 KG 업데이트까지 프로덕션 운영의 핵심을 다룹니다.</p><p><strong>AI / ML</strong> / 20분 읽기 / 시리즈: knowledge-graph</p><p><a href="https://archive.kreathlab.com/tech/knowledge-graph-09-production">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 31 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>knowledge-graph</category>
      <category>ai</category>
      <category>data-engineering</category>
    </item>
    <item>
      <title><![CDATA[8장: 비용 가시성과 FinOps 통합]]></title>
      <link>https://archive.kreathlab.com/tech/platform-engineering-08-finops</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/platform-engineering-08-finops</guid>
      <description><![CDATA[FinOps 원칙과 플랫폼 통합, Backstage 비용 대시보드, 태그 기반 비용 할당, 리소스 생성 시점의 비용 예측, 그리고 AI 기반 비용 최적화를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>FinOps 원칙과 플랫폼 통합, Backstage 비용 대시보드, 태그 기반 비용 할당, 리소스 생성 시점의 비용 예측, 그리고 AI 기반 비용 최적화를 다룹니다.</p><p><strong>인프라</strong> / 16분 읽기 / 시리즈: platform-engineering</p><p><a href="https://archive.kreathlab.com/tech/platform-engineering-08-finops">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 31 Mar 2026 00:00:00 GMT</pubDate>
      <category>infra</category>
      <category>platform-engineering</category>
      <category>devops</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[6장: 데이터 증강 기법]]></title>
      <link>https://archive.kreathlab.com/tech/synthetic-data-06-augmentation</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/synthetic-data-06-augmentation</guid>
      <description><![CDATA[전통적 텍스트 증강부터 LLM 기반 증강, 어려운 예제 생성, 엣지 케이스 증강, 증강 비율 최적화까지 실전 데이터 증강 기법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>전통적 텍스트 증강부터 LLM 기반 증강, 어려운 예제 생성, 엣지 케이스 증강, 증강 비율 최적화까지 실전 데이터 증강 기법을 다룹니다.</p><p><strong>AI / ML</strong> / 21분 읽기 / 시리즈: synthetic-data</p><p><a href="https://archive.kreathlab.com/tech/synthetic-data-06-augmentation">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 31 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>synthetic-data</category>
      <category>ai</category>
      <category>data-engineering</category>
      <category>llm</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[8장: 모델 병렬화와 분산 추론]]></title>
      <link>https://archive.kreathlab.com/tech/llm-inference-08-parallelism</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/llm-inference-08-parallelism</guid>
      <description><![CDATA[텐서 병렬화, 파이프라인 병렬화, 시퀀스 병렬화, Expert 병렬화의 원리를 분석하고, 멀티 GPU 추론 전략과 클러스터 수준 최적화를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>텐서 병렬화, 파이프라인 병렬화, 시퀀스 병렬화, Expert 병렬화의 원리를 분석하고, 멀티 GPU 추론 전략과 클러스터 수준 최적화를 다룹니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: llm-inference</p><p><a href="https://archive.kreathlab.com/tech/llm-inference-08-parallelism">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 30 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>ai</category>
      <category>performance</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[7장: 이벤트 소싱과 CQRS 패턴]]></title>
      <link>https://archive.kreathlab.com/tech/streaming-ai-07-event-sourcing</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/streaming-ai-07-event-sourcing</guid>
      <description><![CDATA[이벤트 소싱과 CQRS 패턴의 원리를 살펴보고, AI 시스템에서의 적용 사례를 다룹니다. 대화 이력 관리, 에이전트 상태 추적, 시간 여행 디버깅, Kafka와 EventStoreDB 활용을 포함합니다.]]></description>
      <content:encoded><![CDATA[<p>이벤트 소싱과 CQRS 패턴의 원리를 살펴보고, AI 시스템에서의 적용 사례를 다룹니다. 대화 이력 관리, 에이전트 상태 추적, 시간 여행 디버깅, Kafka와 EventStoreDB 활용을 포함합니다.</p><p><strong>아키텍처</strong> / 15분 읽기 / 시리즈: streaming-ai</p><p><a href="https://archive.kreathlab.com/tech/streaming-ai-07-event-sourcing">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 30 Mar 2026 00:00:00 GMT</pubDate>
      <category>architecture</category>
      <category>streaming</category>
      <category>ai</category>
    </item>
    <item>
      <title><![CDATA[6장: LLM 기반 데이터 추출 자동화]]></title>
      <link>https://archive.kreathlab.com/tech/structured-output-06-automation</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/structured-output-06-automation</guid>
      <description><![CDATA[대량 문서 처리 파이프라인을 구축하고, 배치 처리, 비동기 추출, 품질 검증 루프, 비용 최적화 전략을 학습합니다.]]></description>
      <content:encoded><![CDATA[<p>대량 문서 처리 파이프라인을 구축하고, 배치 처리, 비동기 추출, 품질 검증 루프, 비용 최적화 전략을 학습합니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: structured-output</p><p><a href="https://archive.kreathlab.com/tech/structured-output-06-automation">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 30 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>structured-output</category>
      <category>ai</category>
      <category>data-engineering</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[5장: Rust에서 Wasm 빌드]]></title>
      <link>https://archive.kreathlab.com/tech/webassembly-05-rust-wasm</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/webassembly-05-rust-wasm</guid>
      <description><![CDATA[Rust에서 WebAssembly를 빌드하는 전체 과정을 다룹니다. wasm-pack, cargo-component, 크기 최적화, WASI 타겟 빌드, 컴포넌트 모델 적용, HTTP 핸들러 실전 예제까지.]]></description>
      <content:encoded><![CDATA[<p>Rust에서 WebAssembly를 빌드하는 전체 과정을 다룹니다. wasm-pack, cargo-component, 크기 최적화, WASI 타겟 빌드, 컴포넌트 모델 적용, HTTP 핸들러 실전 예제까지.</p><p><strong>프로그래밍</strong> / 13분 읽기 / 시리즈: webassembly</p><p><a href="https://archive.kreathlab.com/tech/webassembly-05-rust-wasm">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 30 Mar 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>webassembly</category>
      <category>rust</category>
    </item>
    <item>
      <title><![CDATA[8장: 지식 그래프 쿼리와 추론]]></title>
      <link>https://archive.kreathlab.com/tech/knowledge-graph-08-reasoning</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/knowledge-graph-08-reasoning</guid>
      <description><![CDATA[Cypher 고급 쿼리 패턴, PageRank/커뮤니티 감지/중심성 등 그래프 알고리즘의 실전 활용, LLM과 그래프 추론의 결합, Text2Cypher 자연어 변환까지 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Cypher 고급 쿼리 패턴, PageRank/커뮤니티 감지/중심성 등 그래프 알고리즘의 실전 활용, LLM과 그래프 추론의 결합, Text2Cypher 자연어 변환까지 다룹니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: knowledge-graph</p><p><a href="https://archive.kreathlab.com/tech/knowledge-graph-08-reasoning">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 29 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>knowledge-graph</category>
      <category>ai</category>
      <category>data-engineering</category>
    </item>
    <item>
      <title><![CDATA[7장: 플랫폼 API 설계]]></title>
      <link>https://archive.kreathlab.com/tech/platform-engineering-07-platform-api</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/platform-engineering-07-platform-api</guid>
      <description><![CDATA[Platform as a Product 관점에서의 API 계층 설계, 추상화 수준 결정, 내부 API 버전닝, 인증과 인가, 감사 로깅, 그리고 CLI 도구 제공까지 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Platform as a Product 관점에서의 API 계층 설계, 추상화 수준 결정, 내부 API 버전닝, 인증과 인가, 감사 로깅, 그리고 CLI 도구 제공까지 다룹니다.</p><p><strong>인프라</strong> / 14분 읽기 / 시리즈: platform-engineering</p><p><a href="https://archive.kreathlab.com/tech/platform-engineering-07-platform-api">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 29 Mar 2026 00:00:00 GMT</pubDate>
      <category>infra</category>
      <category>platform-engineering</category>
      <category>devops</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[5장: 데이터 품질 검증과 필터링 파이프라인]]></title>
      <link>https://archive.kreathlab.com/tech/synthetic-data-05-quality-validation</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/synthetic-data-05-quality-validation</guid>
      <description><![CDATA[충실도, 유용성, 프라이버시 3계층 품질 평가 프레임워크와 LLM-as-Judge, 자동 필터링 파이프라인, 중복 제거 전략을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>충실도, 유용성, 프라이버시 3계층 품질 평가 프레임워크와 LLM-as-Judge, 자동 필터링 파이프라인, 중복 제거 전략을 다룹니다.</p><p><strong>AI / ML</strong> / 18분 읽기 / 시리즈: synthetic-data</p><p><a href="https://archive.kreathlab.com/tech/synthetic-data-05-quality-validation">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 29 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>synthetic-data</category>
      <category>ai</category>
      <category>data-engineering</category>
      <category>llm</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[7장: 양자화 추론 — FP8, INT8, INT4]]></title>
      <link>https://archive.kreathlab.com/tech/llm-inference-07-quantization</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/llm-inference-07-quantization</guid>
      <description><![CDATA[양자화의 기초 개념부터 FP8의 부상, W8A8/W4A16 전략, GPTQ/AWQ/SmoothQuant 기법, KV 캐시 양자화까지 정확도와 성능의 트레이드오프를 분석합니다.]]></description>
      <content:encoded><![CDATA[<p>양자화의 기초 개념부터 FP8의 부상, W8A8/W4A16 전략, GPTQ/AWQ/SmoothQuant 기법, KV 캐시 양자화까지 정확도와 성능의 트레이드오프를 분석합니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: llm-inference</p><p><a href="https://archive.kreathlab.com/tech/llm-inference-07-quantization">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 28 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>ai</category>
      <category>performance</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[6장: 실시간 추론 파이프라인 설계]]></title>
      <link>https://archive.kreathlab.com/tech/streaming-ai-06-inference-pipeline</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/streaming-ai-06-inference-pipeline</guid>
      <description><![CDATA[vLLM의 스트리밍 입력, Continuous Batching, 시맨틱 캐싱, 추론 라우터, 멀티모달 실시간 처리 등 백엔드 추론 파이프라인의 핵심 아키텍처를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>vLLM의 스트리밍 입력, Continuous Batching, 시맨틱 캐싱, 추론 라우터, 멀티모달 실시간 처리 등 백엔드 추론 파이프라인의 핵심 아키텍처를 다룹니다.</p><p><strong>아키텍처</strong> / 17분 읽기 / 시리즈: streaming-ai</p><p><a href="https://archive.kreathlab.com/tech/streaming-ai-06-inference-pipeline">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 28 Mar 2026 00:00:00 GMT</pubDate>
      <category>architecture</category>
      <category>streaming</category>
      <category>ai</category>
    </item>
    <item>
      <title><![CDATA[5장: 비정형 데이터에서 구조화된 정보 추출]]></title>
      <link>https://archive.kreathlab.com/tech/structured-output-05-extraction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/structured-output-05-extraction</guid>
      <description><![CDATA[PDF, 이미지, 웹페이지 등 비정형 데이터에서 LLM을 활용하여 구조화된 정보를 추출하는 실전 기법을 학습합니다.]]></description>
      <content:encoded><![CDATA[<p>PDF, 이미지, 웹페이지 등 비정형 데이터에서 LLM을 활용하여 구조화된 정보를 추출하는 실전 기법을 학습합니다.</p><p><strong>AI / ML</strong> / 18분 읽기 / 시리즈: structured-output</p><p><a href="https://archive.kreathlab.com/tech/structured-output-05-extraction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 28 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>structured-output</category>
      <category>ai</category>
      <category>data-engineering</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[4장: 컴포넌트 모델과 WIT]]></title>
      <link>https://archive.kreathlab.com/tech/webassembly-04-component-model</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/webassembly-04-component-model</guid>
      <description><![CDATA[WebAssembly 컴포넌트 모델의 필요성, WIT IDL의 문법과 타입 시스템, 인터페이스와 월드 정의, 컴포넌트 구성을 통한 언어 간 상호운용성을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>WebAssembly 컴포넌트 모델의 필요성, WIT IDL의 문법과 타입 시스템, 인터페이스와 월드 정의, 컴포넌트 구성을 통한 언어 간 상호운용성을 다룹니다.</p><p><strong>프로그래밍</strong> / 13분 읽기 / 시리즈: webassembly</p><p><a href="https://archive.kreathlab.com/tech/webassembly-04-component-model">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 28 Mar 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>webassembly</category>
      <category>rust</category>
    </item>
    <item>
      <title><![CDATA[10장: 프로덕션 하네스 통합 전략]]></title>
      <link>https://archive.kreathlab.com/tech/harness-engineering-10-integration</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/harness-engineering-10-integration</guid>
      <description><![CDATA[전체 하네스 계층 통합, 하네스 성숙도 모델, CI/CD 파이프라인 통합, CLAUDE.md와 AGENTS.md 설계, 팀 협업 전략까지 하네스 엔지니어링의 완결편입니다.]]></description>
      <content:encoded><![CDATA[<p>전체 하네스 계층 통합, 하네스 성숙도 모델, CI/CD 파이프라인 통합, CLAUDE.md와 AGENTS.md 설계, 팀 협업 전략까지 하네스 엔지니어링의 완결편입니다.</p><p><strong>AI / ML</strong> / 22분 읽기 / 시리즈: harness-engineering</p><p><a href="https://archive.kreathlab.com/tech/harness-engineering-10-integration">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 27 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>testing</category>
      <category>evaluation</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[7장: 지식 그래프 임베딩]]></title>
      <link>https://archive.kreathlab.com/tech/knowledge-graph-07-embeddings</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/knowledge-graph-07-embeddings</guid>
      <description><![CDATA[TransE, DistMult, ComplEx 등 관계 예측 모델과 Node2Vec, GraphSAGE 등 노드 임베딩 기법, PyTorch Geometric을 활용한 구현까지 지식 그래프 임베딩의 핵심을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>TransE, DistMult, ComplEx 등 관계 예측 모델과 Node2Vec, GraphSAGE 등 노드 임베딩 기법, PyTorch Geometric을 활용한 구현까지 지식 그래프 임베딩의 핵심을 다룹니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: knowledge-graph</p><p><a href="https://archive.kreathlab.com/tech/knowledge-graph-07-embeddings">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 27 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>knowledge-graph</category>
      <category>ai</category>
      <category>data-engineering</category>
    </item>
    <item>
      <title><![CDATA[6장: 셀프서비스 인프라]]></title>
      <link>https://archive.kreathlab.com/tech/platform-engineering-06-self-service</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/platform-engineering-06-self-service</guid>
      <description><![CDATA[셀프서비스의 핵심 원칙, GitOps 기반 인프라 요청, Crossplane을 활용한 Kubernetes 네이티브 인프라 추상화, 그리고 승인 워크플로우를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>셀프서비스의 핵심 원칙, GitOps 기반 인프라 요청, Crossplane을 활용한 Kubernetes 네이티브 인프라 추상화, 그리고 승인 워크플로우를 다룹니다.</p><p><strong>인프라</strong> / 14분 읽기 / 시리즈: platform-engineering</p><p><a href="https://archive.kreathlab.com/tech/platform-engineering-06-self-service">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 27 Mar 2026 00:00:00 GMT</pubDate>
      <category>infra</category>
      <category>platform-engineering</category>
      <category>devops</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[4장: 구조화된 데이터와 멀티모달 합성]]></title>
      <link>https://archive.kreathlab.com/tech/synthetic-data-04-structured-data</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/synthetic-data-04-structured-data</guid>
      <description><![CDATA[테이블/CSV 합성, JSON/SQL 데이터 생성, 이미지-텍스트 페어 생성, NVIDIA Nemotron 등 멀티모달 합성 데이터 생성 기법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>테이블/CSV 합성, JSON/SQL 데이터 생성, 이미지-텍스트 페어 생성, NVIDIA Nemotron 등 멀티모달 합성 데이터 생성 기법을 다룹니다.</p><p><strong>AI / ML</strong> / 19분 읽기 / 시리즈: synthetic-data</p><p><a href="https://archive.kreathlab.com/tech/synthetic-data-04-structured-data">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 27 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>synthetic-data</category>
      <category>ai</category>
      <category>data-engineering</category>
      <category>llm</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[10장: 실전 프로젝트 -- 컨텍스트 엔지니어링 시스템 구축]]></title>
      <link>https://archive.kreathlab.com/tech/context-engineering-10-project</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/context-engineering-10-project</guid>
      <description><![CDATA[시리즈 전체를 종합하여 실전 컨텍스트 엔지니어링 시스템을 구축합니다. CLAUDE.md 작성, MCP 서버 구현, 멀티에이전트 파이프라인, 측정 대시보드를 실습합니다.]]></description>
      <content:encoded><![CDATA[<p>시리즈 전체를 종합하여 실전 컨텍스트 엔지니어링 시스템을 구축합니다. CLAUDE.md 작성, MCP 서버 구현, 멀티에이전트 파이프라인, 측정 대시보드를 실습합니다.</p><p><strong>AI / ML</strong> / 20분 읽기 / 시리즈: context-engineering</p><p><a href="https://archive.kreathlab.com/tech/context-engineering-10-project">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 26 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>ai-agent</category>
      <category>claude-code</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[6장: Prefix Caching과 프롬프트 최적화]]></title>
      <link>https://archive.kreathlab.com/tech/llm-inference-06-prefix-caching</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/llm-inference-06-prefix-caching</guid>
      <description><![CDATA[시스템 프롬프트 캐싱, Prefix-aware 스케줄링, RadixAttention의 원리를 분석하고, 멀티턴 대화와 평가 워크플로우에서의 성능 개선을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>시스템 프롬프트 캐싱, Prefix-aware 스케줄링, RadixAttention의 원리를 분석하고, 멀티턴 대화와 평가 워크플로우에서의 성능 개선을 다룹니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: llm-inference</p><p><a href="https://archive.kreathlab.com/tech/llm-inference-06-prefix-caching">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 26 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>ai</category>
      <category>performance</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[5장: 스트리밍 LLM 응답 처리]]></title>
      <link>https://archive.kreathlab.com/tech/streaming-ai-05-llm-streaming</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/streaming-ai-05-llm-streaming</guid>
      <description><![CDATA[OpenAI, Anthropic, Google의 스트리밍 API 차이를 비교하고, 구조화된 출력의 파셜 파싱, React 스트리밍 UI 렌더링, Vercel AI SDK 활용법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>OpenAI, Anthropic, Google의 스트리밍 API 차이를 비교하고, 구조화된 출력의 파셜 파싱, React 스트리밍 UI 렌더링, Vercel AI SDK 활용법을 다룹니다.</p><p><strong>아키텍처</strong> / 15분 읽기 / 시리즈: streaming-ai</p><p><a href="https://archive.kreathlab.com/tech/streaming-ai-05-llm-streaming">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 26 Mar 2026 00:00:00 GMT</pubDate>
      <category>architecture</category>
      <category>streaming</category>
      <category>ai</category>
    </item>
    <item>
      <title><![CDATA[4장: Pydantic과 타입 안전 출력]]></title>
      <link>https://archive.kreathlab.com/tech/structured-output-04-pydantic</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/structured-output-04-pydantic</guid>
      <description><![CDATA[Pydantic v2로 LLM 출력 스키마를 정의하고, Instructor 라이브러리로 자동 재시도와 스트리밍 구조화 출력을 구현합니다.]]></description>
      <content:encoded><![CDATA[<p>Pydantic v2로 LLM 출력 스키마를 정의하고, Instructor 라이브러리로 자동 재시도와 스트리밍 구조화 출력을 구현합니다.</p><p><strong>AI / ML</strong> / 15분 읽기 / 시리즈: structured-output</p><p><a href="https://archive.kreathlab.com/tech/structured-output-04-pydantic">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 26 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>structured-output</category>
      <category>ai</category>
      <category>data-engineering</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[3장: WASI — WebAssembly 시스템 인터페이스]]></title>
      <link>https://archive.kreathlab.com/tech/webassembly-03-wasi</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/webassembly-03-wasi</guid>
      <description><![CDATA[WASI의 탄생 배경과 Capability-based 보안 모델, WASI Preview 2의 Worlds 개념, 비동기 지원을 위한 WASI 0.3, 그리고 1.0 로드맵을 상세히 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>WASI의 탄생 배경과 Capability-based 보안 모델, WASI Preview 2의 Worlds 개념, 비동기 지원을 위한 WASI 0.3, 그리고 1.0 로드맵을 상세히 다룹니다.</p><p><strong>프로그래밍</strong> / 14분 읽기 / 시리즈: webassembly</p><p><a href="https://archive.kreathlab.com/tech/webassembly-03-wasi">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 26 Mar 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>webassembly</category>
      <category>rust</category>
    </item>
    <item>
      <title><![CDATA[10장: 실전 프로젝트 — 컨테이너 보안 파이프라인 구축]]></title>
      <link>https://archive.kreathlab.com/tech/container-security-10-project</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/container-security-10-project</guid>
      <description><![CDATA[빌드, 스캔, 서명, 배포, 런타임 모니터링까지 전체 컨테이너 보안 파이프라인을 GitHub Actions와 쿠버네티스 기반으로 통합 구축하는 실전 프로젝트입니다.]]></description>
      <content:encoded><![CDATA[<p>빌드, 스캔, 서명, 배포, 런타임 모니터링까지 전체 컨테이너 보안 파이프라인을 GitHub Actions와 쿠버네티스 기반으로 통합 구축하는 실전 프로젝트입니다.</p><p><strong>인프라</strong> / 18분 읽기 / 시리즈: container-security</p><p><a href="https://archive.kreathlab.com/tech/container-security-10-project">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 25 Mar 2026 00:00:00 GMT</pubDate>
      <category>infra</category>
      <category>security</category>
      <category>kubernetes</category>
      <category>devops</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[9장: 모니터링 하네스 — 프로덕션 관측과 피드백 루프]]></title>
      <link>https://archive.kreathlab.com/tech/harness-engineering-09-monitoring</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/harness-engineering-09-monitoring</guid>
      <description><![CDATA[토큰 사용량, 지연시간, 비용 추적, 드리프트 감지, 품질 모니터링, 알림 설계, 피드백 루프 등 AI 시스템의 관측 가능성 파이프라인을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>토큰 사용량, 지연시간, 비용 추적, 드리프트 감지, 품질 모니터링, 알림 설계, 피드백 루프 등 AI 시스템의 관측 가능성 파이프라인을 다룹니다.</p><p><strong>AI / ML</strong> / 19분 읽기 / 시리즈: harness-engineering</p><p><a href="https://archive.kreathlab.com/tech/harness-engineering-09-monitoring">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 25 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>testing</category>
      <category>evaluation</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[6장: GraphRAG — 그래프 기반 검색 증강 생성]]></title>
      <link>https://archive.kreathlab.com/tech/knowledge-graph-06-graphrag</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/knowledge-graph-06-graphrag</guid>
      <description><![CDATA[Microsoft GraphRAG의 아키텍처, 커뮤니티 요약, 글로벌/로컬 검색 전략, Neo4j GraphRAG Python 라이브러리, 그리고 벡터+그래프+키워드 하이브리드 검색을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Microsoft GraphRAG의 아키텍처, 커뮤니티 요약, 글로벌/로컬 검색 전략, Neo4j GraphRAG Python 라이브러리, 그리고 벡터+그래프+키워드 하이브리드 검색을 다룹니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: knowledge-graph</p><p><a href="https://archive.kreathlab.com/tech/knowledge-graph-06-graphrag">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 25 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>knowledge-graph</category>
      <category>ai</category>
      <category>data-engineering</category>
    </item>
    <item>
      <title><![CDATA[5장: Golden Path 설계와 구현]]></title>
      <link>https://archive.kreathlab.com/tech/platform-engineering-05-golden-path</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/platform-engineering-05-golden-path</guid>
      <description><![CDATA[Golden Path의 정의와 철학, 옵셔널하되 매력적인 경로 설계, Terraform/Pulumi 기반 자동 인프라 프로비저닝, 그리고 80% 이상 자발적 채택률의 비결을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Golden Path의 정의와 철학, 옵셔널하되 매력적인 경로 설계, Terraform/Pulumi 기반 자동 인프라 프로비저닝, 그리고 80% 이상 자발적 채택률의 비결을 다룹니다.</p><p><strong>인프라</strong> / 19분 읽기 / 시리즈: platform-engineering</p><p><a href="https://archive.kreathlab.com/tech/platform-engineering-05-golden-path">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 25 Mar 2026 00:00:00 GMT</pubDate>
      <category>infra</category>
      <category>platform-engineering</category>
      <category>devops</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[3장: 텍스트 데이터 합성 실전]]></title>
      <link>https://archive.kreathlab.com/tech/synthetic-data-03-text-generation</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/synthetic-data-03-text-generation</guid>
      <description><![CDATA[지시-응답 쌍, 대화 데이터, 분류/NER 학습 데이터, 다국어 데이터, 코드 데이터의 합성 파이프라인을 실전 코드와 함께 구축합니다.]]></description>
      <content:encoded><![CDATA[<p>지시-응답 쌍, 대화 데이터, 분류/NER 학습 데이터, 다국어 데이터, 코드 데이터의 합성 파이프라인을 실전 코드와 함께 구축합니다.</p><p><strong>AI / ML</strong> / 21분 읽기 / 시리즈: synthetic-data</p><p><a href="https://archive.kreathlab.com/tech/synthetic-data-03-text-generation">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 25 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>synthetic-data</category>
      <category>ai</category>
      <category>data-engineering</category>
      <category>llm</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[9장: 컨텍스트 품질 측정과 개선]]></title>
      <link>https://archive.kreathlab.com/tech/context-engineering-09-measurement</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/context-engineering-09-measurement</guid>
      <description><![CDATA[컨텍스트 엔지니어링의 효과를 정량적으로 측정하는 방법을 다룹니다. 태스크 성공률, 토큰 효율성, A/B 테스트, 컨텍스트 드리프트 감지를 분석합니다.]]></description>
      <content:encoded><![CDATA[<p>컨텍스트 엔지니어링의 효과를 정량적으로 측정하는 방법을 다룹니다. 태스크 성공률, 토큰 효율성, A/B 테스트, 컨텍스트 드리프트 감지를 분석합니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: context-engineering</p><p><a href="https://archive.kreathlab.com/tech/context-engineering-09-measurement">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 24 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>ai-agent</category>
      <category>claude-code</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[5장: Speculative Decoding — 추측 기반 가속]]></title>
      <link>https://archive.kreathlab.com/tech/llm-inference-05-speculative-decoding</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/llm-inference-05-speculative-decoding</guid>
      <description><![CDATA[Draft-Verify 패러다임으로 자기회귀 디코딩을 가속하는 Speculative Decoding의 원리, 수학적 보장, 그리고 Medusa, Eagle 등 변형 기법을 분석합니다.]]></description>
      <content:encoded><![CDATA[<p>Draft-Verify 패러다임으로 자기회귀 디코딩을 가속하는 Speculative Decoding의 원리, 수학적 보장, 그리고 Medusa, Eagle 등 변형 기법을 분석합니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: llm-inference</p><p><a href="https://archive.kreathlab.com/tech/llm-inference-05-speculative-decoding">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 24 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>ai</category>
      <category>performance</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[4장: gRPC Streaming — 고성능 백엔드 통신]]></title>
      <link>https://archive.kreathlab.com/tech/streaming-ai-04-grpc</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/streaming-ai-04-grpc</guid>
      <description><![CDATA[HTTP/2 기반 gRPC의 4가지 스트리밍 모드, Protobuf 직렬화, 마이크로서비스 간 추론 파이프라인 구현을 다룹니다. gRPC-Web의 제약과 Python/Go 구현 예제를 포함합니다.]]></description>
      <content:encoded><![CDATA[<p>HTTP/2 기반 gRPC의 4가지 스트리밍 모드, Protobuf 직렬화, 마이크로서비스 간 추론 파이프라인 구현을 다룹니다. gRPC-Web의 제약과 Python/Go 구현 예제를 포함합니다.</p><p><strong>아키텍처</strong> / 16분 읽기 / 시리즈: streaming-ai</p><p><a href="https://archive.kreathlab.com/tech/streaming-ai-04-grpc">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 24 Mar 2026 00:00:00 GMT</pubDate>
      <category>architecture</category>
      <category>streaming</category>
      <category>ai</category>
    </item>
    <item>
      <title><![CDATA[3장: 함수 호출(Function Calling)과 도구 사용]]></title>
      <link>https://archive.kreathlab.com/tech/structured-output-03-function-calling</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/structured-output-03-function-calling</guid>
      <description><![CDATA[Function Calling의 원리를 이해하고, OpenAI/Anthropic/Google의 도구 호출 인터페이스로 구조화된 출력을 생성하는 방법을 학습합니다.]]></description>
      <content:encoded><![CDATA[<p>Function Calling의 원리를 이해하고, OpenAI/Anthropic/Google의 도구 호출 인터페이스로 구조화된 출력을 생성하는 방법을 학습합니다.</p><p><strong>AI / ML</strong> / 15분 읽기 / 시리즈: structured-output</p><p><a href="https://archive.kreathlab.com/tech/structured-output-03-function-calling">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 24 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>structured-output</category>
      <category>ai</category>
      <category>data-engineering</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[2장: Wasm 런타임 아키텍처 심층 분석]]></title>
      <link>https://archive.kreathlab.com/tech/webassembly-02-runtime</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/webassembly-02-runtime</guid>
      <description><![CDATA[WebAssembly의 스택 머신 실행 모델, 모듈/인스턴스/메모리/테이블 구조, 주요 런타임(Wasmtime, Wasmer, WasmEdge, V8) 비교, AOT와 JIT 컴파일 전략을 분석합니다.]]></description>
      <content:encoded><![CDATA[<p>WebAssembly의 스택 머신 실행 모델, 모듈/인스턴스/메모리/테이블 구조, 주요 런타임(Wasmtime, Wasmer, WasmEdge, V8) 비교, AOT와 JIT 컴파일 전략을 분석합니다.</p><p><strong>프로그래밍</strong> / 14분 읽기 / 시리즈: webassembly</p><p><a href="https://archive.kreathlab.com/tech/webassembly-02-runtime">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 24 Mar 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>webassembly</category>
      <category>rust</category>
    </item>
    <item>
      <title><![CDATA[10장: 실전 프로젝트 -- LLM 코드 분석 파이프라인 구축]]></title>
      <link>https://archive.kreathlab.com/tech/code-analysis-10-project</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/code-analysis-10-project</guid>
      <description><![CDATA[AST 추출부터 코드 스멜 감지, 리팩터링 제안, 검증, 적용까지 전체 파이프라인을 구축하는 실전 프로젝트입니다. 레거시 프로젝트 현대화 사례와 도입 가이드를 포함합니다.]]></description>
      <content:encoded><![CDATA[<p>AST 추출부터 코드 스멜 감지, 리팩터링 제안, 검증, 적용까지 전체 파이프라인을 구축하는 실전 프로젝트입니다. 레거시 프로젝트 현대화 사례와 도입 가이드를 포함합니다.</p><p><strong>AI / ML</strong> / 22분 읽기 / 시리즈: code-analysis</p><p><a href="https://archive.kreathlab.com/tech/code-analysis-10-project">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 23 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>code-quality</category>
      <category>ai</category>
      <category>llm</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[9장: 공급망 공격 방어와 제로 트러스트]]></title>
      <link>https://archive.kreathlab.com/tech/container-security-09-supply-chain</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/container-security-09-supply-chain</guid>
      <description><![CDATA[의존성 혼동, 타이포스쿼팅, CI 침투 등 공급망 공격 유형을 분석하고, SLSA 프레임워크와 어드미션 컨트롤러 기반의 제로 트러스트 방어 전략을 구축합니다.]]></description>
      <content:encoded><![CDATA[<p>의존성 혼동, 타이포스쿼팅, CI 침투 등 공급망 공격 유형을 분석하고, SLSA 프레임워크와 어드미션 컨트롤러 기반의 제로 트러스트 방어 전략을 구축합니다.</p><p><strong>인프라</strong> / 16분 읽기 / 시리즈: container-security</p><p><a href="https://archive.kreathlab.com/tech/container-security-09-supply-chain">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 23 Mar 2026 00:00:00 GMT</pubDate>
      <category>infra</category>
      <category>security</category>
      <category>kubernetes</category>
      <category>devops</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[8장: 배포 하네스 — 안전한 모델 릴리즈]]></title>
      <link>https://archive.kreathlab.com/tech/harness-engineering-08-deployment</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/harness-engineering-08-deployment</guid>
      <description><![CDATA[카나리 배포, 섀도우 테스팅, A/B 테스트, 블루-그린 배포, 롤백 전략 등 AI 시스템을 프로덕션에 안전하게 배포하는 전략을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>카나리 배포, 섀도우 테스팅, A/B 테스트, 블루-그린 배포, 롤백 전략 등 AI 시스템을 프로덕션에 안전하게 배포하는 전략을 다룹니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: harness-engineering</p><p><a href="https://archive.kreathlab.com/tech/harness-engineering-08-deployment">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 23 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>testing</category>
      <category>evaluation</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[5장: LLM 기반 엔티티 추출과 관계 생성]]></title>
      <link>https://archive.kreathlab.com/tech/knowledge-graph-05-entity-extraction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/knowledge-graph-05-entity-extraction</guid>
      <description><![CDATA[비정형 텍스트에서 LLM을 활용하여 엔티티와 관계를 추출하고, JSON 파싱, 엔티티 해소, Neo4j 적재까지의 전체 파이프라인을 구축하는 방법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>비정형 텍스트에서 LLM을 활용하여 엔티티와 관계를 추출하고, JSON 파싱, 엔티티 해소, Neo4j 적재까지의 전체 파이프라인을 구축하는 방법을 다룹니다.</p><p><strong>AI / ML</strong> / 19분 읽기 / 시리즈: knowledge-graph</p><p><a href="https://archive.kreathlab.com/tech/knowledge-graph-05-entity-extraction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 23 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>knowledge-graph</category>
      <category>ai</category>
      <category>data-engineering</category>
    </item>
    <item>
      <title><![CDATA[4장: 서비스 카탈로그와 소프트웨어 템플릿]]></title>
      <link>https://archive.kreathlab.com/tech/platform-engineering-04-catalog</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/platform-engineering-04-catalog</guid>
      <description><![CDATA[Backstage 소프트웨어 카탈로그의 엔티티 모델, catalog-info.yaml 스키마, 메타데이터 표준화, 그리고 Scaffolder를 활용한 프로젝트 자동 생성을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Backstage 소프트웨어 카탈로그의 엔티티 모델, catalog-info.yaml 스키마, 메타데이터 표준화, 그리고 Scaffolder를 활용한 프로젝트 자동 생성을 다룹니다.</p><p><strong>인프라</strong> / 15분 읽기 / 시리즈: platform-engineering</p><p><a href="https://archive.kreathlab.com/tech/platform-engineering-04-catalog">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 23 Mar 2026 00:00:00 GMT</pubDate>
      <category>infra</category>
      <category>platform-engineering</category>
      <category>devops</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[2장: LLM 기반 데이터 생성의 원리와 기법]]></title>
      <link>https://archive.kreathlab.com/tech/synthetic-data-02-llm-generation</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/synthetic-data-02-llm-generation</guid>
      <description><![CDATA[프롬프트 기반 생성, 디스틸레이션, Evol-Instruct, Self-Instruct 등 LLM을 활용한 합성 데이터 생성 기법의 원리와 실전 적용법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>프롬프트 기반 생성, 디스틸레이션, Evol-Instruct, Self-Instruct 등 LLM을 활용한 합성 데이터 생성 기법의 원리와 실전 적용법을 다룹니다.</p><p><strong>AI / ML</strong> / 21분 읽기 / 시리즈: synthetic-data</p><p><a href="https://archive.kreathlab.com/tech/synthetic-data-02-llm-generation">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 23 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>synthetic-data</category>
      <category>ai</category>
      <category>data-engineering</category>
      <category>llm</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[10장: 실전 프로젝트 -- AI 테스트 자동화 파이프라인]]></title>
      <link>https://archive.kreathlab.com/tech/ai-testing-10-project</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-testing-10-project</guid>
      <description><![CDATA[단위, 통합, E2E, 시각적, 변이 테스트를 하나의 AI 테스트 자동화 파이프라인으로 통합합니다. Codium, Playwright, Applitools를 결합한 CI/CD 파이프라인과 대시보드, 도입 로드맵, ROI 측정을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>단위, 통합, E2E, 시각적, 변이 테스트를 하나의 AI 테스트 자동화 파이프라인으로 통합합니다. Codium, Playwright, Applitools를 결합한 CI/CD 파이프라인과 대시보드, 도입 로드맵, ROI 측정을 다룹니다.</p><p><strong>AI / ML</strong> / 23분 읽기 / 시리즈: ai-testing</p><p><a href="https://archive.kreathlab.com/tech/ai-testing-10-project">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 22 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>testing</category>
      <category>automation</category>
      <category>quality-assurance</category>
      <category>ai</category>
    </item>
    <item>
      <title><![CDATA[8장: 컨텍스트 엔지니어링 도구와 기법]]></title>
      <link>https://archive.kreathlab.com/tech/context-engineering-08-tools</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/context-engineering-08-tools</guid>
      <description><![CDATA[Claude Code, Cursor, GitHub Copilot에서 컨텍스트를 최적화하는 구체적 방법과, MCP 서버를 통한 동적 컨텍스트 확장 기법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Claude Code, Cursor, GitHub Copilot에서 컨텍스트를 최적화하는 구체적 방법과, MCP 서버를 통한 동적 컨텍스트 확장 기법을 다룹니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: context-engineering</p><p><a href="https://archive.kreathlab.com/tech/context-engineering-08-tools">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 22 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>ai-agent</category>
      <category>claude-code</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[4장: Continuous Batching과 동적 배칭]]></title>
      <link>https://archive.kreathlab.com/tech/llm-inference-04-continuous-batching</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/llm-inference-04-continuous-batching</guid>
      <description><![CDATA[정적 배칭의 한계를 분석하고, Continuous Batching의 iteration-level 스케줄링 원리와 vLLM, TGI, TensorRT-LLM의 구현 차이를 비교합니다.]]></description>
      <content:encoded><![CDATA[<p>정적 배칭의 한계를 분석하고, Continuous Batching의 iteration-level 스케줄링 원리와 vLLM, TGI, TensorRT-LLM의 구현 차이를 비교합니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: llm-inference</p><p><a href="https://archive.kreathlab.com/tech/llm-inference-04-continuous-batching">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 22 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>ai</category>
      <category>performance</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[3장: WebSocket — 양방향 실시간 통신]]></title>
      <link>https://archive.kreathlab.com/tech/streaming-ai-03-websocket</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/streaming-ai-03-websocket</guid>
      <description><![CDATA[WebSocket의 핸드셰이크, 프레이밍 구조, 양방향 통신의 강점과 상태 관리의 복잡성을 분석합니다. AI 채팅에서의 생성 중단, Socket.IO, 스케일링 전략을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>WebSocket의 핸드셰이크, 프레이밍 구조, 양방향 통신의 강점과 상태 관리의 복잡성을 분석합니다. AI 채팅에서의 생성 중단, Socket.IO, 스케일링 전략을 다룹니다.</p><p><strong>아키텍처</strong> / 17분 읽기 / 시리즈: streaming-ai</p><p><a href="https://archive.kreathlab.com/tech/streaming-ai-03-websocket">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 22 Mar 2026 00:00:00 GMT</pubDate>
      <category>architecture</category>
      <category>streaming</category>
      <category>ai</category>
    </item>
    <item>
      <title><![CDATA[2장: JSON Schema 기반 LLM 출력 제어]]></title>
      <link>https://archive.kreathlab.com/tech/structured-output-02-json-schema</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/structured-output-02-json-schema</guid>
      <description><![CDATA[JSON Schema 기초 문법을 학습하고, OpenAI, Anthropic, Google 주요 프로바이더의 구조화된 출력 API를 실습합니다.]]></description>
      <content:encoded><![CDATA[<p>JSON Schema 기초 문법을 학습하고, OpenAI, Anthropic, Google 주요 프로바이더의 구조화된 출력 API를 실습합니다.</p><p><strong>AI / ML</strong> / 13분 읽기 / 시리즈: structured-output</p><p><a href="https://archive.kreathlab.com/tech/structured-output-02-json-schema">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 22 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>structured-output</category>
      <category>ai</category>
      <category>data-engineering</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[1장: WebAssembly의 등장과 핵심 개념]]></title>
      <link>https://archive.kreathlab.com/tech/webassembly-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/webassembly-01-introduction</guid>
      <description><![CDATA[WebAssembly란 무엇인지, 바이너리 포맷과 텍스트 포맷의 차이, 선형 메모리 모델과 샌드박스 보안, 그리고 2026년 현재 Wasm 생태계의 전체 지도를 살펴봅니다.]]></description>
      <content:encoded><![CDATA[<p>WebAssembly란 무엇인지, 바이너리 포맷과 텍스트 포맷의 차이, 선형 메모리 모델과 샌드박스 보안, 그리고 2026년 현재 Wasm 생태계의 전체 지도를 살펴봅니다.</p><p><strong>프로그래밍</strong> / 14분 읽기 / 시리즈: webassembly</p><p><a href="https://archive.kreathlab.com/tech/webassembly-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 22 Mar 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>webassembly</category>
      <category>rust</category>
    </item>
    <item>
      <title><![CDATA[9장: CI/CD 통합과 지속적 코드 품질 관리]]></title>
      <link>https://archive.kreathlab.com/tech/code-analysis-09-pipeline</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/code-analysis-09-pipeline</guid>
      <description><![CDATA[LLM 기반 코드 분석을 CI/CD 파이프라인에 통합하는 방법을 학습합니다. PR별 자동 분석, 품질 게이트, 기술 부채 대시보드와 GitHub Actions 구축을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>LLM 기반 코드 분석을 CI/CD 파이프라인에 통합하는 방법을 학습합니다. PR별 자동 분석, 품질 게이트, 기술 부채 대시보드와 GitHub Actions 구축을 다룹니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: code-analysis</p><p><a href="https://archive.kreathlab.com/tech/code-analysis-09-pipeline">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 21 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>code-quality</category>
      <category>ai</category>
      <category>llm</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[8장: 시크릿 관리]]></title>
      <link>https://archive.kreathlab.com/tech/container-security-08-secrets</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/container-security-08-secrets</guid>
      <description><![CDATA[쿠버네티스 Secrets의 한계를 이해하고, HashiCorp Vault, External Secrets Operator, Sealed Secrets로 안전한 시크릿 관리 체계를 구축합니다.]]></description>
      <content:encoded><![CDATA[<p>쿠버네티스 Secrets의 한계를 이해하고, HashiCorp Vault, External Secrets Operator, Sealed Secrets로 안전한 시크릿 관리 체계를 구축합니다.</p><p><strong>인프라</strong> / 14분 읽기 / 시리즈: container-security</p><p><a href="https://archive.kreathlab.com/tech/container-security-08-secrets">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 21 Mar 2026 00:00:00 GMT</pubDate>
      <category>infra</category>
      <category>security</category>
      <category>kubernetes</category>
      <category>devops</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[7장: 오케스트레이션 하네스 — 워크플로우 제어]]></title>
      <link>https://archive.kreathlab.com/tech/harness-engineering-07-orchestration</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/harness-engineering-07-orchestration</guid>
      <description><![CDATA[에이전트 라이프사이클 관리, 도구 오케스트레이션, 서브에이전트 관리, 상태 관리, 에러 복구 등 복잡한 AI 워크플로우를 조율하는 방법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>에이전트 라이프사이클 관리, 도구 오케스트레이션, 서브에이전트 관리, 상태 관리, 에러 복구 등 복잡한 AI 워크플로우를 조율하는 방법을 다룹니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: harness-engineering</p><p><a href="https://archive.kreathlab.com/tech/harness-engineering-07-orchestration">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 21 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>testing</category>
      <category>evaluation</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[4장: Amazon Neptune과 기타 그래프 DB]]></title>
      <link>https://archive.kreathlab.com/tech/knowledge-graph-04-neptune</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/knowledge-graph-04-neptune</guid>
      <description><![CDATA[Amazon Neptune의 아키텍처와 Bedrock 통합, 그리고 TigerGraph, JanusGraph, Memgraph 등 주요 그래프 데이터베이스를 비교하며 프로젝트에 맞는 선택 가이드를 제공합니다.]]></description>
      <content:encoded><![CDATA[<p>Amazon Neptune의 아키텍처와 Bedrock 통합, 그리고 TigerGraph, JanusGraph, Memgraph 등 주요 그래프 데이터베이스를 비교하며 프로젝트에 맞는 선택 가이드를 제공합니다.</p><p><strong>AI / ML</strong> / 13분 읽기 / 시리즈: knowledge-graph</p><p><a href="https://archive.kreathlab.com/tech/knowledge-graph-04-neptune">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 21 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>knowledge-graph</category>
      <category>ai</category>
      <category>data-engineering</category>
    </item>
    <item>
      <title><![CDATA[3장: Backstage로 개발자 포털 구축]]></title>
      <link>https://archive.kreathlab.com/tech/platform-engineering-03-backstage</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/platform-engineering-03-backstage</guid>
      <description><![CDATA[Spotify 오픈소스이자 CNCF 졸업 프로젝트인 Backstage의 아키텍처, 설치, 소프트웨어 카탈로그, TechDocs, 플러그인 시스템을 실습합니다.]]></description>
      <content:encoded><![CDATA[<p>Spotify 오픈소스이자 CNCF 졸업 프로젝트인 Backstage의 아키텍처, 설치, 소프트웨어 카탈로그, TechDocs, 플러그인 시스템을 실습합니다.</p><p><strong>인프라</strong> / 16분 읽기 / 시리즈: platform-engineering</p><p><a href="https://archive.kreathlab.com/tech/platform-engineering-03-backstage">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 21 Mar 2026 00:00:00 GMT</pubDate>
      <category>infra</category>
      <category>platform-engineering</category>
      <category>devops</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[1장: 합성 데이터의 부상과 엔지니어링]]></title>
      <link>https://archive.kreathlab.com/tech/synthetic-data-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/synthetic-data-01-introduction</guid>
      <description><![CDATA[합성 데이터의 정의와 유형, 2026년 합성 데이터가 필수가 된 배경, 데이터 순환 문제와 생태계 전체 지도를 살펴봅니다.]]></description>
      <content:encoded><![CDATA[<p>합성 데이터의 정의와 유형, 2026년 합성 데이터가 필수가 된 배경, 데이터 순환 문제와 생태계 전체 지도를 살펴봅니다.</p><p><strong>AI / ML</strong> / 19분 읽기 / 시리즈: synthetic-data</p><p><a href="https://archive.kreathlab.com/tech/synthetic-data-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 21 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>synthetic-data</category>
      <category>ai</category>
      <category>data-engineering</category>
      <category>llm</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[10장: CI/CD 통합과 품질 게이트 구축]]></title>
      <link>https://archive.kreathlab.com/tech/ai-eval-harness-10-cicd-integration</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-eval-harness-10-cicd-integration</guid>
      <description><![CDATA[GitHub Actions에 평가 파이프라인을 통합하고, 품질 게이트를 설계하고, 회귀 테스트를 자동화합니다. 프롬��트 변경 감지, 드리프트 모니터링까지 종합 평가 CI/CD 파이프라인을 구축합니다.]]></description>
      <content:encoded><![CDATA[<p>GitHub Actions에 평가 파이프라인을 통합하고, 품질 게이트를 설계하고, 회귀 테스트를 자동화합니다. 프롬��트 변경 감지, 드리프트 모니터링까지 종합 평가 CI/CD 파이프라인을 구축합니다.</p><p><strong>AI / ML</strong> / 18분 읽기 / 시리즈: ai-eval-harness</p><p><a href="https://archive.kreathlab.com/tech/ai-eval-harness-10-cicd-integration">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 20 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>evaluation</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[9장: Agentic QA -- 자율 테스트 에이전트]]></title>
      <link>https://archive.kreathlab.com/tech/ai-testing-09-agentic-qa</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-testing-09-agentic-qa</guid>
      <description><![CDATA[유저 스토리에서 Gherkin 시나리오를 거쳐 실행 가능한 테스트로 자동 변환하는 Agentic QA의 아키텍처, 자율 탐색 테스트, Human-on-the-loop 감독 체계, 그리고 품질 게이트 통합을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>유저 스토리에서 Gherkin 시나리오를 거쳐 실행 가능한 테스트로 자동 변환하는 Agentic QA의 아키텍처, 자율 탐색 테스트, Human-on-the-loop 감독 체계, 그리고 품질 게이트 통합을 다룹니다.</p><p><strong>AI / ML</strong> / 22분 읽기 / 시리즈: ai-testing</p><p><a href="https://archive.kreathlab.com/tech/ai-testing-09-agentic-qa">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 20 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>testing</category>
      <category>automation</category>
      <category>quality-assurance</category>
      <category>ai</category>
    </item>
    <item>
      <title><![CDATA[7장: 컨텍스트 격리와 멀티에이전트 설계]]></title>
      <link>https://archive.kreathlab.com/tech/context-engineering-07-isolation</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/context-engineering-07-isolation</guid>
      <description><![CDATA[멀티에이전트 시스템에서 에이전트별 컨텍스트를 격리하고, 크로스 오염을 방지하며, 공유 컨텍스트를 효과적으로 관리하는 전략을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>멀티에이전트 시스템에서 에이전트별 컨텍스트를 격리하고, 크로스 오염을 방지하며, 공유 컨텍스트를 효과적으로 관리하는 전략을 다룹니다.</p><p><strong>AI / ML</strong> / 18분 읽기 / 시리즈: context-engineering</p><p><a href="https://archive.kreathlab.com/tech/context-engineering-07-isolation">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 20 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>ai-agent</category>
      <category>claude-code</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[3장: PagedAttention과 vLLM]]></title>
      <link>https://archive.kreathlab.com/tech/llm-inference-03-pagedattention</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/llm-inference-03-pagedattention</guid>
      <description><![CDATA[OS 가상 메모리에서 영감받은 PagedAttention의 원리를 설명하고, vLLM의 아키텍처와 Automatic Prefix Caching, 계층적 KV 캐시를 분석합니다.]]></description>
      <content:encoded><![CDATA[<p>OS 가상 메모리에서 영감받은 PagedAttention의 원리를 설명하고, vLLM의 아키텍처와 Automatic Prefix Caching, 계층적 KV 캐시를 분석합니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: llm-inference</p><p><a href="https://archive.kreathlab.com/tech/llm-inference-03-pagedattention">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 20 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>ai</category>
      <category>performance</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[2장: SSE(Server-Sent Events) 심층 분석]]></title>
      <link>https://archive.kreathlab.com/tech/streaming-ai-02-sse</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/streaming-ai-02-sse</guid>
      <description><![CDATA[HTTP 기반 단방향 스트리밍 프로토콜인 SSE의 동작 원리, EventSource API, 자동 재연결 메커니즘을 분석하고 Next.js와 FastAPI에서의 LLM 토큰 스트리밍 구현을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>HTTP 기반 단방향 스트리밍 프로토콜인 SSE의 동작 원리, EventSource API, 자동 재연결 메커니즘을 분석하고 Next.js와 FastAPI에서의 LLM 토큰 스트리밍 구현을 다룹니다.</p><p><strong>아키텍처</strong> / 14분 읽기 / 시리즈: streaming-ai</p><p><a href="https://archive.kreathlab.com/tech/streaming-ai-02-sse">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 20 Mar 2026 00:00:00 GMT</pubDate>
      <category>architecture</category>
      <category>streaming</category>
      <category>ai</category>
    </item>
    <item>
      <title><![CDATA[1장: Structured Output의 필요성과 핵심 개념]]></title>
      <link>https://archive.kreathlab.com/tech/structured-output-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/structured-output-01-introduction</guid>
      <description><![CDATA[LLM 비정형 출력의 한계를 분석하고, 구조화된 출력의 3가지 접근 방식과 제약 디코딩의 원리를 살펴봅니다.]]></description>
      <content:encoded><![CDATA[<p>LLM 비정형 출력의 한계를 분석하고, 구조화된 출력의 3가지 접근 방식과 제약 디코딩의 원리를 살펴봅니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: structured-output</p><p><a href="https://archive.kreathlab.com/tech/structured-output-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 20 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>structured-output</category>
      <category>ai</category>
      <category>data-engineering</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[10장: 실전 프로젝트 -- Agentic Workflow 시스템 구축]]></title>
      <link>https://archive.kreathlab.com/tech/agentic-workflow-10-project</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/agentic-workflow-10-project</guid>
      <description><![CDATA[고객 지원 자동화 워크플로우를 LangGraph로 구현하고, HITL, 감사 로깅, 엔터프라이즈 통합, 보안까지 포함한 프로덕션 수준 시스템을 구축합니다.]]></description>
      <content:encoded><![CDATA[<p>고객 지원 자동화 워크플로우를 LangGraph로 구현하고, HITL, 감사 로깅, 엔터프라이즈 통합, 보안까지 포함한 프로덕션 수준 시스템을 구축합니다.</p><p><strong>AI / ML</strong> / 22분 읽기 / 시리즈: agentic-workflow</p><p><a href="https://archive.kreathlab.com/tech/agentic-workflow-10-project">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 19 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>workflow</category>
      <category>ai</category>
      <category>automation</category>
    </item>
    <item>
      <title><![CDATA[8장: 아키텍처 분석과 시각화]]></title>
      <link>https://archive.kreathlab.com/tech/code-analysis-08-architecture</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/code-analysis-08-architecture</guid>
      <description><![CDATA[LLM을 활용한 아키텍처 분석, 순환 의존성 감지, 레이어 위반 탐지, 마이크로서비스 경계 제안과 아키텍처 다이어그램 자동 생성을 학습합니다.]]></description>
      <content:encoded><![CDATA[<p>LLM을 활용한 아키텍처 분석, 순환 의존성 감지, 레이어 위반 탐지, 마이크로서비스 경계 제안과 아키텍처 다이어그램 자동 생성을 학습합니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: code-analysis</p><p><a href="https://archive.kreathlab.com/tech/code-analysis-08-architecture">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 19 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>code-quality</category>
      <category>ai</category>
      <category>llm</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[7장: 네트워크 정책과 서비스 메시]]></title>
      <link>https://archive.kreathlab.com/tech/container-security-07-network</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/container-security-07-network</guid>
      <description><![CDATA[쿠버네티스 NetworkPolicy로 기본 거부 정책을 구현하고, Calico/Cilium 네트워크 정책과 Istio mTLS로 컨테이너 간 통신을 안전하게 제어합니다.]]></description>
      <content:encoded><![CDATA[<p>쿠버네티스 NetworkPolicy로 기본 거부 정책을 구현하고, Calico/Cilium 네트워크 정책과 Istio mTLS로 컨테이너 간 통신을 안전하게 제어합니다.</p><p><strong>인프라</strong> / 14분 읽기 / 시리즈: container-security</p><p><a href="https://archive.kreathlab.com/tech/container-security-07-network">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 19 Mar 2026 00:00:00 GMT</pubDate>
      <category>infra</category>
      <category>security</category>
      <category>kubernetes</category>
      <category>devops</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[6장: 가드레일 하네스 — 안전 장치 설계와 구현]]></title>
      <link>https://archive.kreathlab.com/tech/harness-engineering-06-guardrail-harness</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/harness-engineering-06-guardrail-harness</guid>
      <description><![CDATA[프롬프트 인젝션 방어, 유해 콘텐츠 필터링, Guardrails AI와 NeMo Guardrails 프레임워크, 다계층 방어 전략을 통해 AI 시스템의 안전을 보장하는 방법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>프롬프트 인젝션 방어, 유해 콘텐츠 필터링, Guardrails AI와 NeMo Guardrails 프레임워크, 다계층 방어 전략을 통해 AI 시스템의 안전을 보장하는 방법을 다룹니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: harness-engineering</p><p><a href="https://archive.kreathlab.com/tech/harness-engineering-06-guardrail-harness">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 19 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>testing</category>
      <category>evaluation</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[3장: Neo4j — 프로퍼티 그래프 데이터베이스]]></title>
      <link>https://archive.kreathlab.com/tech/knowledge-graph-03-neo4j</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/knowledge-graph-03-neo4j</guid>
      <description><![CDATA[Neo4j의 아키텍처, Cypher 쿼리 언어, 벡터 인덱스, GDS 라이브러리, Python 드라이버까지 지식 그래프 구축에 필요한 Neo4j의 핵심 기능을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Neo4j의 아키텍처, Cypher 쿼리 언어, 벡터 인덱스, GDS 라이브러리, Python 드라이버까지 지식 그래프 구축에 필요한 Neo4j의 핵심 기능을 다룹니다.</p><p><strong>AI / ML</strong> / 14분 읽기 / 시리즈: knowledge-graph</p><p><a href="https://archive.kreathlab.com/tech/knowledge-graph-03-neo4j">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 19 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>knowledge-graph</category>
      <category>ai</category>
      <category>data-engineering</category>
    </item>
    <item>
      <title><![CDATA[2장: 내부 개발자 플랫폼(IDP) 설계]]></title>
      <link>https://archive.kreathlab.com/tech/platform-engineering-02-idp-design</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/platform-engineering-02-idp-design</guid>
      <description><![CDATA[IDP 아키텍처의 핵심 구성 요소, 사용자 리서치 기반 설계, Build vs Buy 의사결정, 그리고 MVP부터 점진적 확장까지의 전략을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>IDP 아키텍처의 핵심 구성 요소, 사용자 리서치 기반 설계, Build vs Buy 의사결정, 그리고 MVP부터 점진적 확장까지의 전략을 다룹니다.</p><p><strong>인프라</strong> / 20분 읽기 / 시리즈: platform-engineering</p><p><a href="https://archive.kreathlab.com/tech/platform-engineering-02-idp-design">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 19 Mar 2026 00:00:00 GMT</pubDate>
      <category>infra</category>
      <category>platform-engineering</category>
      <category>devops</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[10장: 실전 프로젝트 — 에이전트 품질 보증 파이프라인]]></title>
      <link>https://archive.kreathlab.com/tech/agent-testing-10-project</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/agent-testing-10-project</guid>
      <description><![CDATA[단위 테스트부터 프로덕션 모니터링까지 전체 에이전트 품질 보증 파이프라인을 구축하고, 도입 체크리스트와 성숙도 모델을 제시합니다.]]></description>
      <content:encoded><![CDATA[<p>단위 테스트부터 프로덕션 모니터링까지 전체 에이전트 품질 보증 파이프라인을 구축하고, 도입 체크리스트와 성숙도 모델을 제시합니다.</p><p><strong>AI / ML</strong> / 19분 읽기 / 시리즈: agent-testing</p><p><a href="https://archive.kreathlab.com/tech/agent-testing-10-project">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 18 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>testing</category>
      <category>ai</category>
      <category>evaluation</category>
      <category>quality-assurance</category>
    </item>
    <item>
      <title><![CDATA[9장: 자동화된 모델 비교 파이프라인]]></title>
      <link>https://archive.kreathlab.com/tech/ai-eval-harness-09-comparison-pipeline</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-eval-harness-09-comparison-pipeline</guid>
      <description><![CDATA[ELO 레이팅과 리더보드 구현, A/B 테스트 자동화, 비용/지연시간/품질 트레이드오프 분석, 모델 선택 자동화, 비교 리포트 자동 생성까지 모델 비교 파이프라인을 구축합니다.]]></description>
      <content:encoded><![CDATA[<p>ELO 레이팅과 리더보드 구현, A/B 테스트 자동화, 비용/지연시간/품질 트레이드오프 분석, 모델 선택 자동화, 비교 리포트 자동 생성까지 모델 비교 파이프라인을 구축합니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: ai-eval-harness</p><p><a href="https://archive.kreathlab.com/tech/ai-eval-harness-09-comparison-pipeline">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 18 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>evaluation</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[8장: AI QA 파이프라인 구축]]></title>
      <link>https://archive.kreathlab.com/tech/ai-testing-08-pipeline</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-testing-08-pipeline</guid>
      <description><![CDATA[변경 영향 분석 기반 테스트 선택, 위험 기반 우선순위, 플레이키 테스트 자동 격리, 병렬 실행 최적화, 결함 예측, GitHub Actions/GitLab CI 통합을 다루는 AI QA 파이프라인 구축 가이드입니다.]]></description>
      <content:encoded><![CDATA[<p>변경 영향 분석 기반 테스트 선택, 위험 기반 우선순위, 플레이키 테스트 자동 격리, 병렬 실행 최적화, 결함 예측, GitHub Actions/GitLab CI 통합을 다루는 AI QA 파이프라인 구축 가이드입니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: ai-testing</p><p><a href="https://archive.kreathlab.com/tech/ai-testing-08-pipeline">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 18 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>testing</category>
      <category>automation</category>
      <category>quality-assurance</category>
      <category>ai</category>
    </item>
    <item>
      <title><![CDATA[6장: 컨텍스트 정렬과 포맷 최적화]]></title>
      <link>https://archive.kreathlab.com/tech/context-engineering-06-ordering-format</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/context-engineering-06-ordering-format</guid>
      <description><![CDATA[어텐션 메커니즘과 위치 편향을 이해하고, 정보 배치 전략과 XML/마크다운/JSON 포맷 비교를 통해 컨텍스트 구조를 최적화합니다.]]></description>
      <content:encoded><![CDATA[<p>어텐션 메커니즘과 위치 편향을 이해하고, 정보 배치 전략과 XML/마크다운/JSON 포맷 비교를 통해 컨텍스트 구조를 최적화합니다.</p><p><strong>AI / ML</strong> / 15분 읽기 / 시리즈: context-engineering</p><p><a href="https://archive.kreathlab.com/tech/context-engineering-06-ordering-format">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 18 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>ai-agent</category>
      <category>claude-code</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[2장: KV 캐시 메커니즘 심층 분석]]></title>
      <link>https://archive.kreathlab.com/tech/llm-inference-02-kv-cache</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/llm-inference-02-kv-cache</guid>
      <description><![CDATA[트랜스포머 Attention에서 KV 캐시의 역할과 메모리 사용량 계산법을 다루고, MQA/GQA 등 캐시 절감 기법과 압축 전략을 분석합니다.]]></description>
      <content:encoded><![CDATA[<p>트랜스포머 Attention에서 KV 캐시의 역할과 메모리 사용량 계산법을 다루고, MQA/GQA 등 캐시 절감 기법과 압축 전략을 분석합니다.</p><p><strong>AI / ML</strong> / 15분 읽기 / 시리즈: llm-inference</p><p><a href="https://archive.kreathlab.com/tech/llm-inference-02-kv-cache">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 18 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>ai</category>
      <category>performance</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[1장: 스트리밍 아키텍처의 필요성과 핵심 개념]]></title>
      <link>https://archive.kreathlab.com/tech/streaming-ai-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/streaming-ai-01-introduction</guid>
      <description><![CDATA[요청-응답 모델의 한계를 넘어 스트리밍 아키텍처가 왜 AI 시대의 필수 인프라인지 살펴봅니다. TTFT, TPOT 등 핵심 지표와 프로토콜 생태계를 개관합니다.]]></description>
      <content:encoded><![CDATA[<p>요청-응답 모델의 한계를 넘어 스트리밍 아키텍처가 왜 AI 시대의 필수 인프라인지 살펴봅니다. TTFT, TPOT 등 핵심 지표와 프로토콜 생태계를 개관합니다.</p><p><strong>아키텍처</strong> / 15분 읽기 / 시리즈: streaming-ai</p><p><a href="https://archive.kreathlab.com/tech/streaming-ai-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 18 Mar 2026 00:00:00 GMT</pubDate>
      <category>architecture</category>
      <category>streaming</category>
      <category>ai</category>
    </item>
    <item>
      <title><![CDATA[9장: 보안과 거버넌스]]></title>
      <link>https://archive.kreathlab.com/tech/agentic-workflow-09-governance</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/agentic-workflow-09-governance</guid>
      <description><![CDATA[최소 권한 원칙, 도구별 권한 제어, 비밀 관리, 입출력 검증, 비용 제어, 에이전트 거버넌스 프레임워크, 위험 평가, 모니터링과 알림 전략을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>최소 권한 원칙, 도구별 권한 제어, 비밀 관리, 입출력 검증, 비용 제어, 에이전트 거버넌스 프레임워크, 위험 평가, 모니터링과 알림 전략을 다룹니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: agentic-workflow</p><p><a href="https://archive.kreathlab.com/tech/agentic-workflow-09-governance">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 17 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>workflow</category>
      <category>ai</category>
      <category>automation</category>
    </item>
    <item>
      <title><![CDATA[7장: 보안 취약점 분석과 자동 수정]]></title>
      <link>https://archive.kreathlab.com/tech/code-analysis-07-security</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/code-analysis-07-security</guid>
      <description><![CDATA[SAST와 LLM을 결합한 보안 취약점 탐지, OWASP Top 10 자동 검출, 취약점 자동 수정 제안과 CI/CD 보안 게이트 구축을 학습합니다.]]></description>
      <content:encoded><![CDATA[<p>SAST와 LLM을 결합한 보안 취약점 탐지, OWASP Top 10 자동 검출, 취약점 자동 수정 제안과 CI/CD 보안 게이트 구축을 학습합니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: code-analysis</p><p><a href="https://archive.kreathlab.com/tech/code-analysis-07-security">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 17 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>code-quality</category>
      <category>ai</category>
      <category>llm</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[6장: 런타임 보안 — Falco와 위협 감지]]></title>
      <link>https://archive.kreathlab.com/tech/container-security-06-runtime</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/container-security-06-runtime</guid>
      <description><![CDATA[Falco의 eBPF 기반 시스콜 모니터링으로 컨테이너 런타임 위협을 실시간 감지하고, 규칙 작성부터 알림 통합까지 런타임 보안 체계를 구축합니다.]]></description>
      <content:encoded><![CDATA[<p>Falco의 eBPF 기반 시스콜 모니터링으로 컨테이너 런타임 위협을 실시간 감지하고, 규칙 작성부터 알림 통합까지 런타임 보안 체계를 구축합니다.</p><p><strong>인프라</strong> / 14분 읽기 / 시리즈: container-security</p><p><a href="https://archive.kreathlab.com/tech/container-security-06-runtime">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 17 Mar 2026 00:00:00 GMT</pubDate>
      <category>infra</category>
      <category>security</category>
      <category>kubernetes</category>
      <category>devops</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[5장: 평가 하네스 — 모델 성능 측정 파이프라인]]></title>
      <link>https://archive.kreathlab.com/tech/harness-engineering-05-evaluation-harness</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/harness-engineering-05-evaluation-harness</guid>
      <description><![CDATA[lm-evaluation-harness, Inspect AI, HELM 프레임워크 분석과 커스텀 평가 하네스 설계, 벤치마크 스위트 구성, 자동화된 모델 비교 방법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>lm-evaluation-harness, Inspect AI, HELM 프레임워크 분석과 커스텀 평가 하네스 설계, 벤치마크 스위트 구성, 자동화된 모델 비교 방법을 다룹니다.</p><p><strong>AI / ML</strong> / 14분 읽기 / 시리즈: harness-engineering</p><p><a href="https://archive.kreathlab.com/tech/harness-engineering-05-evaluation-harness">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 17 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>testing</category>
      <category>evaluation</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[2장: 그래프 데이터 모델링 기초]]></title>
      <link>https://archive.kreathlab.com/tech/knowledge-graph-02-modeling</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/knowledge-graph-02-modeling</guid>
      <description><![CDATA[프로퍼티 그래프와 RDF의 차이, 노드/엣지/속성 설계 원칙, 온톨로지 설계부터 실전 도메인 모델링까지 지식 그래프의 데이터 모델링 기초를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>프로퍼티 그래프와 RDF의 차이, 노드/엣지/속성 설계 원칙, 온톨로지 설계부터 실전 도메인 모델링까지 지식 그래프의 데이터 모델링 기초를 다룹니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: knowledge-graph</p><p><a href="https://archive.kreathlab.com/tech/knowledge-graph-02-modeling">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 17 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>knowledge-graph</category>
      <category>ai</category>
      <category>data-engineering</category>
    </item>
    <item>
      <title><![CDATA[1장: Platform Engineering의 등장과 핵심 개념]]></title>
      <link>https://archive.kreathlab.com/tech/platform-engineering-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/platform-engineering-01-introduction</guid>
      <description><![CDATA[DevOps의 한계에서 출발한 Platform Engineering의 등장 배경, 인지 부하 감소를 위한 내부 개발자 플랫폼(IDP)의 정의, 그리고 2026년 현황과 트렌드를 살펴봅니다.]]></description>
      <content:encoded><![CDATA[<p>DevOps의 한계에서 출발한 Platform Engineering의 등장 배경, 인지 부하 감소를 위한 내부 개발자 플랫폼(IDP)의 정의, 그리고 2026년 현황과 트렌드를 살펴봅니다.</p><p><strong>인프라</strong> / 15분 읽기 / 시리즈: platform-engineering</p><p><a href="https://archive.kreathlab.com/tech/platform-engineering-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 17 Mar 2026 00:00:00 GMT</pubDate>
      <category>infra</category>
      <category>platform-engineering</category>
      <category>devops</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[10장: 실전 프로젝트 — 에이전트 메모리 시스템 구축]]></title>
      <link>https://archive.kreathlab.com/tech/agent-memory-10-project</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/agent-memory-10-project</guid>
      <description><![CDATA[Mem0와 Zep을 활용한 듀얼 레이어 메모리 시스템 구축, 메모리 압축 파이프라인, 성능 벤치마킹, 프로덕션 운영 체크리스트까지 실전 가이드를 제공합니다.]]></description>
      <content:encoded><![CDATA[<p>Mem0와 Zep을 활용한 듀얼 레이어 메모리 시스템 구축, 메모리 압축 파이프라인, 성능 벤치마킹, 프로덕션 운영 체크리스트까지 실전 가이드를 제공합니다.</p><p><strong>AI / ML</strong> / 19분 읽기 / 시리즈: agent-memory</p><p><a href="https://archive.kreathlab.com/tech/agent-memory-10-project">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 16 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>llm</category>
      <category>ai-agent</category>
    </item>
    <item>
      <title><![CDATA[9장: CI/CD 통합과 품질 게이트]]></title>
      <link>https://archive.kreathlab.com/tech/agent-testing-09-cicd</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/agent-testing-09-cicd</guid>
      <description><![CDATA[GitHub Actions에서 에이전트 테스트를 실행하고, 품질 게이트 임계값을 설계하며, PR별 평가와 온라인 평가를 연결하는 자동화 전략을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>GitHub Actions에서 에이전트 테스트를 실행하고, 품질 게이트 임계값을 설계하며, PR별 평가와 온라인 평가를 연결하는 자동화 전략을 다룹니다.</p><p><strong>AI / ML</strong> / 18분 읽기 / 시리즈: agent-testing</p><p><a href="https://archive.kreathlab.com/tech/agent-testing-09-cicd">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 16 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>testing</category>
      <category>ai</category>
      <category>evaluation</category>
      <category>quality-assurance</category>
    </item>
    <item>
      <title><![CDATA[8장: 벤치마크 스위트 설계 원칙과 실전]]></title>
      <link>https://archive.kreathlab.com/tech/ai-eval-harness-08-benchmark-suite</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-eval-harness-08-benchmark-suite</guid>
      <description><![CDATA[벤치마크 오염 문제, 좋은 벤치마크의 조건, 다차원 평가 설계, 도메인별 벤치마크 구축, 데이터셋 버전 관리, 통계적 유의성 검증까지 벤치마크 스위트 설계의 전체를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>벤치마크 오염 문제, 좋은 벤치마크의 조건, 다차원 평가 설계, 도메인별 벤치마크 구축, 데이터셋 버전 관리, 통계적 유의성 검증까지 벤치마크 스위트 설계의 전체를 다룹니다.</p><p><strong>AI / ML</strong> / 18분 읽기 / 시리즈: ai-eval-harness</p><p><a href="https://archive.kreathlab.com/tech/ai-eval-harness-08-benchmark-suite">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 16 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>evaluation</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[7장: 테스트 유지보수 자동화]]></title>
      <link>https://archive.kreathlab.com/tech/ai-testing-07-maintenance</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-testing-07-maintenance</guid>
      <description><![CDATA[테스트 로트(Test Rot) 문제의 근본 원인과 AI 기반 셀프 힐링, 셀렉터 자동 재바인딩, 테스트 코드 리팩터링, 중복 테스트 감지, 커버리지 갭 분석 등 유지보수 비용 절감 전략을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>테스트 로트(Test Rot) 문제의 근본 원인과 AI 기반 셀프 힐링, 셀렉터 자동 재바인딩, 테스트 코드 리팩터링, 중복 테스트 감지, 커버리지 갭 분석 등 유지보수 비용 절감 전략을 다룹니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: ai-testing</p><p><a href="https://archive.kreathlab.com/tech/ai-testing-07-maintenance">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 16 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>testing</category>
      <category>automation</category>
      <category>quality-assurance</category>
      <category>ai</category>
    </item>
    <item>
      <title><![CDATA[5장: 컨텍스트 압축과 정보 밀도 최적화]]></title>
      <link>https://archive.kreathlab.com/tech/context-engineering-05-compression</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/context-engineering-05-compression</guid>
      <description><![CDATA[토큰 한계와 비용을 최적화하면서 핵심 의미를 보존하는 컨텍스트 압축 기법을 다룹니다. 코드 요약, 인터페이스 추출, 트리 구조 압축 등을 분석합니다.]]></description>
      <content:encoded><![CDATA[<p>토큰 한계와 비용을 최적화하면서 핵심 의미를 보존하는 컨텍스트 압축 기법을 다룹니다. 코드 요약, 인터페이스 추출, 트리 구조 압축 등을 분석합니다.</p><p><strong>AI / ML</strong> / 15분 읽기 / 시리즈: context-engineering</p><p><a href="https://archive.kreathlab.com/tech/context-engineering-05-compression">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 16 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>ai-agent</category>
      <category>claude-code</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[1장: LLM 추론의 기초와 병목 지점]]></title>
      <link>https://archive.kreathlab.com/tech/llm-inference-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/llm-inference-01-introduction</guid>
      <description><![CDATA[트랜스포머 기반 LLM의 추론 과정을 Prefill과 Decode 단계로 나누어 분석하고, 메모리 바운드와 컴퓨트 바운드의 개념, 핵심 지연시간 지표를 정리합니다.]]></description>
      <content:encoded><![CDATA[<p>트랜스포머 기반 LLM의 추론 과정을 Prefill과 Decode 단계로 나누어 분석하고, 메모리 바운드와 컴퓨트 바운드의 개념, 핵심 지연시간 지표를 정리합니다.</p><p><strong>AI / ML</strong> / 14분 읽기 / 시리즈: llm-inference</p><p><a href="https://archive.kreathlab.com/tech/llm-inference-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 16 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>ai</category>
      <category>performance</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[8장: 엔터프라이즈 시스템 통합]]></title>
      <link>https://archive.kreathlab.com/tech/agentic-workflow-08-integration</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/agentic-workflow-08-integration</guid>
      <description><![CDATA[ERP/CRM/ITSM 연동, MCP 기반 도구 통합, API 게이트웨이, 이벤트 드리븐 통합, 레거시 시스템 어댑터, 트랜잭션 경계 설계를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>ERP/CRM/ITSM 연동, MCP 기반 도구 통합, API 게이트웨이, 이벤트 드리븐 통합, 레거시 시스템 어댑터, 트랜잭션 경계 설계를 다룹니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: agentic-workflow</p><p><a href="https://archive.kreathlab.com/tech/agentic-workflow-08-integration">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 15 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>workflow</category>
      <category>ai</category>
      <category>automation</category>
    </item>
    <item>
      <title><![CDATA[6장: 코드 마이그레이션 자동화]]></title>
      <link>https://archive.kreathlab.com/tech/code-analysis-06-migration</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/code-analysis-06-migration</guid>
      <description><![CDATA[LLM을 활용한 언어/프레임워크 마이그레이션 자동화를 학습합니다. Java에서 Kotlin, React Class에서 Hooks로의 전환과 의미 보존 검증 기법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>LLM을 활용한 언어/프레임워크 마이그레이션 자동화를 학습합니다. Java에서 Kotlin, React Class에서 Hooks로의 전환과 의미 보존 검증 기법을 다룹니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: code-analysis</p><p><a href="https://archive.kreathlab.com/tech/code-analysis-06-migration">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 15 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>code-quality</category>
      <category>ai</category>
      <category>llm</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[5장: Sigstore와 Cosign — 이미지 서명]]></title>
      <link>https://archive.kreathlab.com/tech/container-security-05-signing</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/container-security-05-signing</guid>
      <description><![CDATA[Sigstore 에코시스템(Cosign, Fulcio, Rekor)으로 컨테이너 이미지에 키리스 서명을 적용하고, SLSA 프레임워크 기반의 빌드 출처 증명을 구현합니다.]]></description>
      <content:encoded><![CDATA[<p>Sigstore 에코시스템(Cosign, Fulcio, Rekor)으로 컨테이너 이미지에 키리스 서명을 적용하고, SLSA 프레임워크 기반의 빌드 출처 증명을 구현합니다.</p><p><strong>인프라</strong> / 15분 읽기 / 시리즈: container-security</p><p><a href="https://archive.kreathlab.com/tech/container-security-05-signing">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 15 Mar 2026 00:00:00 GMT</pubDate>
      <category>infra</category>
      <category>security</category>
      <category>kubernetes</category>
      <category>devops</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[4장: 테스트 하네스 — AI 시스템의 품질 보증]]></title>
      <link>https://archive.kreathlab.com/tech/harness-engineering-04-test-harness</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/harness-engineering-04-test-harness</guid>
      <description><![CDATA[비결정적 출력 테스트, 스냅샷 테스트, 속성 기반 테스트, 회귀 테스트, 에이전트 행동 테스트 등 AI 시스템 테스트의 핵심 기법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>비결정적 출력 테스트, 스냅샷 테스트, 속성 기반 테스트, 회귀 테스트, 에이전트 행동 테스트 등 AI 시스템 테스트의 핵심 기법을 다룹니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: harness-engineering</p><p><a href="https://archive.kreathlab.com/tech/harness-engineering-04-test-harness">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 15 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>testing</category>
      <category>evaluation</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[1장: Knowledge Graph의 등장과 AI에서의 역할]]></title>
      <link>https://archive.kreathlab.com/tech/knowledge-graph-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/knowledge-graph-01-introduction</guid>
      <description><![CDATA[지식 그래프의 정의와 역사, 벡터 검색의 한계를 그래프가 어떻게 보완하는지, GraphRAG의 35% 정확도 향상 사례까지 Knowledge Graph와 AI 결합의 전체 그림을 소개합니다.]]></description>
      <content:encoded><![CDATA[<p>지식 그래프의 정의와 역사, 벡터 검색의 한계를 그래프가 어떻게 보완하는지, GraphRAG의 35% 정확도 향상 사례까지 Knowledge Graph와 AI 결합의 전체 그림을 소개합니다.</p><p><strong>AI / ML</strong> / 19분 읽기 / 시리즈: knowledge-graph</p><p><a href="https://archive.kreathlab.com/tech/knowledge-graph-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 15 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>knowledge-graph</category>
      <category>ai</category>
      <category>data-engineering</category>
    </item>
    <item>
      <title><![CDATA[10장: 실전 프로젝트 — 온디바이스 AI 시스템 구축]]></title>
      <link>https://archive.kreathlab.com/tech/on-device-ai-10-practical-project</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/on-device-ai-10-practical-project</guid>
      <description><![CDATA[시리즈 전체의 기법을 종합하여 프라이버시 보존 문서 분석 시스템을 구축합니다. 로컬 LLM, 로컬 임베딩, 로컬 벡터 DB로 완전한 오프라인 AI를 실현합니다.]]></description>
      <content:encoded><![CDATA[<p>시리즈 전체의 기법을 종합하여 프라이버시 보존 문서 분석 시스템을 구축합니다. 로컬 LLM, 로컬 임베딩, 로컬 벡터 DB로 완전한 오프라인 AI를 실현합니다.</p><p><strong>AI / ML</strong> / 15분 읽기 / 시리즈: on-device-ai</p><p><a href="https://archive.kreathlab.com/tech/on-device-ai-10-practical-project">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 15 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>performance</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[9장: 프로젝트 메모리와 코딩 에이전트]]></title>
      <link>https://archive.kreathlab.com/tech/agent-memory-09-project-memory</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/agent-memory-09-project-memory</guid>
      <description><![CDATA[CLAUDE.md 기반 프로젝트 메모리, 코드베이스 컨텍스트 지속, 세션 간 학습, 팀 메모리 설계 패턴 등 코딩 에이전트에 특화된 메모리 시스템을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>CLAUDE.md 기반 프로젝트 메모리, 코드베이스 컨텍스트 지속, 세션 간 학습, 팀 메모리 설계 패턴 등 코딩 에이전트에 특화된 메모리 시스템을 다룹니다.</p><p><strong>AI / ML</strong> / 18분 읽기 / 시리즈: agent-memory</p><p><a href="https://archive.kreathlab.com/tech/agent-memory-09-project-memory">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 14 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>llm</category>
      <category>ai-agent</category>
    </item>
    <item>
      <title><![CDATA[8장: 에이전트 시뮬레이션과 레드티밍]]></title>
      <link>https://archive.kreathlab.com/tech/agent-testing-08-simulation</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/agent-testing-08-simulation</guid>
      <description><![CDATA[시뮬레이션 사용자 기반 적대적 테스트, 엣지 케이스 자동 생성, 스트레스 테스트, 안전성 가드레일 검증, 자동 레드티밍 기법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>시뮬레이션 사용자 기반 적대적 테스트, 엣지 케이스 자동 생성, 스트레스 테스트, 안전성 가드레일 검증, 자동 레드티밍 기법을 다룹니다.</p><p><strong>AI / ML</strong> / 19분 읽기 / 시리즈: agent-testing</p><p><a href="https://archive.kreathlab.com/tech/agent-testing-08-simulation">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 14 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>testing</category>
      <category>ai</category>
      <category>evaluation</category>
      <category>quality-assurance</category>
    </item>
    <item>
      <title><![CDATA[7장: 커스텀 평가 하네스 설계와 구축]]></title>
      <link>https://archive.kreathlab.com/tech/ai-eval-harness-07-custom-harness</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-eval-harness-07-custom-harness</guid>
      <description><![CDATA[도메인 특화 평가 하네스를 처음부터 설계하고 구축합니다. 평가 태스크 설계, 메트릭 정의, LLM-as-Judge 구현, 인간 평가 통합, Golden Dataset 관리를 코드와 함께 실습합니다.]]></description>
      <content:encoded><![CDATA[<p>도메인 특화 평가 하네스를 처음부터 설계하고 구축합니다. 평가 태스크 설계, 메트릭 정의, LLM-as-Judge 구현, 인간 평가 통합, Golden Dataset 관리를 코드와 함께 실습합니다.</p><p><strong>AI / ML</strong> / 19분 읽기 / 시리즈: ai-eval-harness</p><p><a href="https://archive.kreathlab.com/tech/ai-eval-harness-07-custom-harness">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 14 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>evaluation</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[10장: 실전 프로젝트 — 보안 강화 LLM 애플리케이션]]></title>
      <link>https://archive.kreathlab.com/tech/ai-security-10-practical-project</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-security-10-practical-project</guid>
      <description><![CDATA[시리즈 전체의 보안 기법을 종합하여 프로덕션 수준의 보안 강화 LLM 애플리케이션을 설계하고 구현합니다. 다층 방어, 가드레일, 모니터링을 통합한 실전 시스템입니다.]]></description>
      <content:encoded><![CDATA[<p>시리즈 전체의 보안 기법을 종합하여 프로덕션 수준의 보안 강화 LLM 애플리케이션을 설계하고 구현합니다. 다층 방어, 가드레일, 모니터링을 통합한 실전 시스템입니다.</p><p><strong>AI / ML</strong> / 13분 읽기 / 시리즈: ai-security</p><p><a href="https://archive.kreathlab.com/tech/ai-security-10-practical-project">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 14 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>testing</category>
      <category>security</category>
    </item>
    <item>
      <title><![CDATA[6장: 시각적 회귀 테스트]]></title>
      <link>https://archive.kreathlab.com/tech/ai-testing-06-visual</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-testing-06-visual</guid>
      <description><![CDATA[픽셀 비교의 한계를 넘어 Visual AI 기반 시각적 회귀 테스트를 다룹니다. Applitools Eyes, Percy, Chromatic 비교 분석과 동적 콘텐츠 처리, 반응형 레이아웃 테스트, 스토리북 통합을 안내합니다.]]></description>
      <content:encoded><![CDATA[<p>픽셀 비교의 한계를 넘어 Visual AI 기반 시각적 회귀 테스트를 다룹니다. Applitools Eyes, Percy, Chromatic 비교 분석과 동적 콘텐츠 처리, 반응형 레이아웃 테스트, 스토리북 통합을 안내합니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: ai-testing</p><p><a href="https://archive.kreathlab.com/tech/ai-testing-06-visual">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 14 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>testing</category>
      <category>automation</category>
      <category>quality-assurance</category>
      <category>ai</category>
    </item>
    <item>
      <title><![CDATA[4장: 컨텍스트 선택과 검색 전략]]></title>
      <link>https://archive.kreathlab.com/tech/context-engineering-04-selection</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/context-engineering-04-selection</guid>
      <description><![CDATA[코드베이스에서 관련 파일을 정밀하게 선택하는 기법을 다룹니다. @-멘션 시스템, RAG 기반 코드 검색, 의존성 그래프 추적, 변경 영향 분석을 분석합니다.]]></description>
      <content:encoded><![CDATA[<p>코드베이스에서 관련 파일을 정밀하게 선택하는 기법을 다룹니다. @-멘션 시스템, RAG 기반 코드 검색, 의존성 그래프 추적, 변경 영향 분석을 분석합니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: context-engineering</p><p><a href="https://archive.kreathlab.com/tech/context-engineering-04-selection">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 14 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>ai-agent</category>
      <category>claude-code</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[7장: 감사 로깅과 컴플라이언스]]></title>
      <link>https://archive.kreathlab.com/tech/agentic-workflow-07-audit</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/agentic-workflow-07-audit</guid>
      <description><![CDATA[에이전트 행동 추적, 불변 감사 로그 설계, 규제 요구사항 대응, 설명 가능성, 재현 가능성, OpenTelemetry 통합, 보존 정책을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>에이전트 행동 추적, 불변 감사 로그 설계, 규제 요구사항 대응, 설명 가능성, 재현 가능성, OpenTelemetry 통합, 보존 정책을 다룹니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: agentic-workflow</p><p><a href="https://archive.kreathlab.com/tech/agentic-workflow-07-audit">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 13 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>workflow</category>
      <category>ai</category>
      <category>automation</category>
    </item>
    <item>
      <title><![CDATA[10장: 실전 프로젝트 — AI-Native 시스템 아키텍처 설계]]></title>
      <link>https://archive.kreathlab.com/tech/ai-architecture-10-practical-project</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-architecture-10-practical-project</guid>
      <description><![CDATA[시리즈 전체의 아키텍처 패턴을 종합하여 프로덕션 AI-Native 시스템을 설계합니다. 전체 아키텍처 다이어그램, 기술 선택, 배포 전략을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>시리즈 전체의 아키텍처 패턴을 종합하여 프로덕션 AI-Native 시스템을 설계합니다. 전체 아키텍처 다이어그램, 기술 선택, 배포 전략을 다룹니다.</p><p><strong>아키텍처</strong> / 17분 읽기 / 시리즈: ai-architecture</p><p><a href="https://archive.kreathlab.com/tech/ai-architecture-10-practical-project">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 13 Mar 2026 00:00:00 GMT</pubDate>
      <category>architecture</category>
      <category>architecture</category>
      <category>llm</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[5장: LLM 기반 자동 리팩터링]]></title>
      <link>https://archive.kreathlab.com/tech/code-analysis-05-refactoring</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/code-analysis-05-refactoring</guid>
      <description><![CDATA[LLM을 활용한 자동 리팩터링의 패턴, 멀티에이전트 아키텍처, 검증 파이프라인을 학습합니다. 37%에서 98%로 정밀도를 끌어올리는 실전 기법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>LLM을 활용한 자동 리팩터링의 패턴, 멀티에이전트 아키텍처, 검증 파이프라인을 학습합니다. 37%에서 98%로 정밀도를 끌어올리는 실전 기법을 다룹니다.</p><p><strong>AI / ML</strong> / 15분 읽기 / 시리즈: code-analysis</p><p><a href="https://archive.kreathlab.com/tech/code-analysis-05-refactoring">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 13 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>code-quality</category>
      <category>ai</category>
      <category>llm</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[4장: SBOM 생성과 관리]]></title>
      <link>https://archive.kreathlab.com/tech/container-security-04-sbom</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/container-security-04-sbom</guid>
      <description><![CDATA[SBOM(소프트웨어 자재 명세서)의 개념과 필요성, SPDX와 CycloneDX 형식을 비교하고, Syft와 Trivy로 SBOM을 생성하여 공급망 가시성을 확보하는 방법을 실습합니다.]]></description>
      <content:encoded><![CDATA[<p>SBOM(소프트웨어 자재 명세서)의 개념과 필요성, SPDX와 CycloneDX 형식을 비교하고, Syft와 Trivy로 SBOM을 생성하여 공급망 가시성을 확보하는 방법을 실습합니다.</p><p><strong>인프라</strong> / 14분 읽기 / 시리즈: container-security</p><p><a href="https://archive.kreathlab.com/tech/container-security-04-sbom">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 13 Mar 2026 00:00:00 GMT</pubDate>
      <category>infra</category>
      <category>security</category>
      <category>kubernetes</category>
      <category>devops</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[3장: AI 모델 래핑과 입출력 제어]]></title>
      <link>https://archive.kreathlab.com/tech/harness-engineering-03-model-wrapping</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/harness-engineering-03-model-wrapping</guid>
      <description><![CDATA[모델 추상화 계층 설계, 프롬프트 구성과 컨텍스트 주입, 스키마 기반 출력 제어, 폴백 전략 등 AI 모델의 입출력을 체계적으로 관리하는 방법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>모델 추상화 계층 설계, 프롬프트 구성과 컨텍스트 주입, 스키마 기반 출력 제어, 폴백 전략 등 AI 모델의 입출력을 체계적으로 관리하는 방법을 다룹니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: harness-engineering</p><p><a href="https://archive.kreathlab.com/tech/harness-engineering-03-model-wrapping">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 13 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>testing</category>
      <category>evaluation</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[9장: 성능 벤치마킹과 최적화]]></title>
      <link>https://archive.kreathlab.com/tech/on-device-ai-09-benchmarking</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/on-device-ai-09-benchmarking</guid>
      <description><![CDATA[온디바이스 AI 시스템의 성능 벤치마킹 방법론, 핵심 지표, 하드웨어별 성능 비교, 그리고 토큰 처리량과 메모리 사용을 최적화하는 기법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>온디바이스 AI 시스템의 성능 벤치마킹 방법론, 핵심 지표, 하드웨어별 성능 비교, 그리고 토큰 처리량과 메모리 사용을 최적화하는 기법을 다룹니다.</p><p><strong>AI / ML</strong> / 13분 읽기 / 시리즈: on-device-ai</p><p><a href="https://archive.kreathlab.com/tech/on-device-ai-09-benchmarking">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 13 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>performance</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[8장: 프로덕션 듀얼 레이어 아키텍처]]></title>
      <link>https://archive.kreathlab.com/tech/agent-memory-08-dual-layer</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/agent-memory-08-dual-layer</guid>
      <description><![CDATA[Hot Path와 Cold Path를 결합한 듀얼 레이어 메모리 아키텍처의 설계, 하이브리드 검색, 메모리 라우팅, 비용-지연시간 최적화 전략을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Hot Path와 Cold Path를 결합한 듀얼 레이어 메모리 아키텍처의 설계, 하이브리드 검색, 메모리 라우팅, 비용-지연시간 최적화 전략을 다룹니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: agent-memory</p><p><a href="https://archive.kreathlab.com/tech/agent-memory-08-dual-layer">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 12 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>llm</category>
      <category>ai-agent</category>
    </item>
    <item>
      <title><![CDATA[7장: 에이전트 안정성 메트릭]]></title>
      <link>https://archive.kreathlab.com/tech/agent-testing-07-stability</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/agent-testing-07-stability</guid>
      <description><![CDATA[태스크 성공률 추이, 행동 드리프트 감지, 응답 길이 변동, 지연시간 안정성, 비용 변동성 등 에이전트의 장기적 안정성을 추적하는 메트릭과 대시보드 설계를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>태스크 성공률 추이, 행동 드리프트 감지, 응답 길이 변동, 지연시간 안정성, 비용 변동성 등 에이전트의 장기적 안정성을 추적하는 메트릭과 대시보드 설계를 다룹니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: agent-testing</p><p><a href="https://archive.kreathlab.com/tech/agent-testing-07-stability">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 12 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>testing</category>
      <category>ai</category>
      <category>evaluation</category>
      <category>quality-assurance</category>
    </item>
    <item>
      <title><![CDATA[6장: 평가 도구 생태계 비교와 선택 기준]]></title>
      <link>https://archive.kreathlab.com/tech/ai-eval-harness-06-ecosystem</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-eval-harness-06-ecosystem</guid>
      <description><![CDATA[DeepEval, promptfoo, Evidently AI, W&B Weave, LangSmith, Ragas 등 실무 평가 도구를 비교합니다. 학술 vs 실무 평가의 차이점과 프레임워크 선택 의사결정 트리를 제시합니다.]]></description>
      <content:encoded><![CDATA[<p>DeepEval, promptfoo, Evidently AI, W&amp;B Weave, LangSmith, Ragas 등 실무 평가 도구를 비교합니다. 학술 vs 실무 평가의 차이점과 프레임워크 선택 의사결정 트리를 제시합니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: ai-eval-harness</p><p><a href="https://archive.kreathlab.com/tech/ai-eval-harness-06-ecosystem">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 12 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>evaluation</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[9장: 보안 모니터링과 사고 대응]]></title>
      <link>https://archive.kreathlab.com/tech/ai-security-09-monitoring</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-security-09-monitoring</guid>
      <description><![CDATA[LLM 시스템의 보안 모니터링 아키텍처, 이상 탐지, 보안 대시보드, 사고 대응 프로세스, 그리고 지속적 보안 운영 체계를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>LLM 시스템의 보안 모니터링 아키텍처, 이상 탐지, 보안 대시보드, 사고 대응 프로세스, 그리고 지속적 보안 운영 체계를 다룹니다.</p><p><strong>AI / ML</strong> / 10분 읽기 / 시리즈: ai-security</p><p><a href="https://archive.kreathlab.com/tech/ai-security-09-monitoring">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 12 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>testing</category>
      <category>security</category>
    </item>
    <item>
      <title><![CDATA[5장: 변이 테스트(Mutation Testing)]]></title>
      <link>https://archive.kreathlab.com/tech/ai-testing-05-mutation</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-testing-05-mutation</guid>
      <description><![CDATA[변이 테스트의 원리와 변이 연산자를 이해하고, Stryker, PIT, mutmut 도구로 AI 생성 테스트의 품질을 검증하는 방법을 다룹니다. 변이 점수 측정과 비용-효과 분석도 포함합니다.]]></description>
      <content:encoded><![CDATA[<p>변이 테스트의 원리와 변이 연산자를 이해하고, Stryker, PIT, mutmut 도구로 AI 생성 테스트의 품질을 검증하는 방법을 다룹니다. 변이 점수 측정과 비용-효과 분석도 포함합니다.</p><p><strong>AI / ML</strong> / 15분 읽기 / 시리즈: ai-testing</p><p><a href="https://archive.kreathlab.com/tech/ai-testing-05-mutation">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 12 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>testing</category>
      <category>automation</category>
      <category>quality-assurance</category>
      <category>ai</category>
    </item>
    <item>
      <title><![CDATA[3장: CLAUDE.md와 AGENTS.md 고급 전략]]></title>
      <link>https://archive.kreathlab.com/tech/context-engineering-03-context-files</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/context-engineering-03-context-files</guid>
      <description><![CDATA[CLAUDE.md와 AGENTS.md 컨텍스트 파일의 설계 원칙, 효과적인 구조화 방법, 컨텍스트 블로트의 위험성과 대응 전략을 심층 분석합니다.]]></description>
      <content:encoded><![CDATA[<p>CLAUDE.md와 AGENTS.md 컨텍스트 파일의 설계 원칙, 효과적인 구조화 방법, 컨텍스트 블로트의 위험성과 대응 전략을 심층 분석합니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: context-engineering</p><p><a href="https://archive.kreathlab.com/tech/context-engineering-03-context-files">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 12 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>ai-agent</category>
      <category>claude-code</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[6장: 상태 관리와 체크포인팅]]></title>
      <link>https://archive.kreathlab.com/tech/agentic-workflow-06-state</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/agentic-workflow-06-state</guid>
      <description><![CDATA[Agentic Workflow의 상태 모델, 이벤트 소싱, 체크포인트 저장소 선택, 멱등성 보장, 상태 복원과 버전 마이그레이션, 분산 상태 일관성 전략을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Agentic Workflow의 상태 모델, 이벤트 소싱, 체크포인트 저장소 선택, 멱등성 보장, 상태 복원과 버전 마이그레이션, 분산 상태 일관성 전략을 다룹니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: agentic-workflow</p><p><a href="https://archive.kreathlab.com/tech/agentic-workflow-06-state">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 11 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>workflow</category>
      <category>ai</category>
      <category>automation</category>
    </item>
    <item>
      <title><![CDATA[9장: 확장성과 멀티테넌시 설계]]></title>
      <link>https://archive.kreathlab.com/tech/ai-architecture-09-scalability</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-architecture-09-scalability</guid>
      <description><![CDATA[AI 시스템의 수평 확장, 멀티테넌시 아키텍처, 속도 제한, 공정 스케줄링, 그리고 대규모 AI 서비스 운영을 위한 인프라 설계를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>AI 시스템의 수평 확장, 멀티테넌시 아키텍처, 속도 제한, 공정 스케줄링, 그리고 대규모 AI 서비스 운영을 위한 인프라 설계를 다룹니다.</p><p><strong>아키텍처</strong> / 17분 읽기 / 시리즈: ai-architecture</p><p><a href="https://archive.kreathlab.com/tech/ai-architecture-09-scalability">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 11 Mar 2026 00:00:00 GMT</pubDate>
      <category>architecture</category>
      <category>architecture</category>
      <category>llm</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[4장: 코드 스멜 감지와 기술 부채 정량화]]></title>
      <link>https://archive.kreathlab.com/tech/code-analysis-04-code-smells</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/code-analysis-04-code-smells</guid>
      <description><![CDATA[LLM 기반 코드 스멜 탐지와 CodeScene Code Health 메트릭을 활용한 기술 부채 정량화를 학습합니다. 우선순위 기반 리팩터링 계획 수립까지 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>LLM 기반 코드 스멜 탐지와 CodeScene Code Health 메트릭을 활용한 기술 부채 정량화를 학습합니다. 우선순위 기반 리팩터링 계획 수립까지 다룹니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: code-analysis</p><p><a href="https://archive.kreathlab.com/tech/code-analysis-04-code-smells">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 11 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>code-quality</category>
      <category>ai</category>
      <category>llm</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[3장: 이미지 스캐닝과 취약점 관리]]></title>
      <link>https://archive.kreathlab.com/tech/container-security-03-scanning</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/container-security-03-scanning</guid>
      <description><![CDATA[Trivy, Grype, Snyk 컨테이너 스캐너를 비교하고, CI/CD 파이프라인에 취약점 스캐닝 게이트를 통합하여 안전한 이미지만 배포하는 방법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Trivy, Grype, Snyk 컨테이너 스캐너를 비교하고, CI/CD 파이프라인에 취약점 스캐닝 게이트를 통합하여 안전한 이미지만 배포하는 방법을 다룹니다.</p><p><strong>인프라</strong> / 15분 읽기 / 시리즈: container-security</p><p><a href="https://archive.kreathlab.com/tech/container-security-03-scanning">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 11 Mar 2026 00:00:00 GMT</pubDate>
      <category>infra</category>
      <category>security</category>
      <category>kubernetes</category>
      <category>devops</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[2장: 하네스 아키텍처 설계 패턴]]></title>
      <link>https://archive.kreathlab.com/tech/harness-engineering-02-architecture</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/harness-engineering-02-architecture</guid>
      <description><![CDATA[래핑, 미들웨어, 사이드카, 파이프라인, 이벤트 기반 등 AI 시스템 하네스의 5가지 핵심 아키텍처 패턴과 적용 시나리오를 분석합니다.]]></description>
      <content:encoded><![CDATA[<p>래핑, 미들웨어, 사이드카, 파이프라인, 이벤트 기반 등 AI 시스템 하네스의 5가지 핵심 아키텍처 패턴과 적용 시나리오를 분석합니다.</p><p><strong>AI / ML</strong> / 21분 읽기 / 시리즈: harness-engineering</p><p><a href="https://archive.kreathlab.com/tech/harness-engineering-02-architecture">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 11 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>testing</category>
      <category>evaluation</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[8장: 온디바이스 AI 애플리케이션 패턴]]></title>
      <link>https://archive.kreathlab.com/tech/on-device-ai-08-application-patterns</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/on-device-ai-08-application-patterns</guid>
      <description><![CDATA[온디바이스 AI를 활용한 실전 애플리케이션 설계 패턴 — 하이브리드 추론, 오프라인 우선, 프라이버시 보존, 개인화 학습, 그리고 에지-클라우드 협업을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>온디바이스 AI를 활용한 실전 애플리케이션 설계 패턴 — 하이브리드 추론, 오프라인 우선, 프라이버시 보존, 개인화 학습, 그리고 에지-클라우드 협업을 다룹니다.</p><p><strong>AI / ML</strong> / 13분 읽기 / 시리즈: on-device-ai</p><p><a href="https://archive.kreathlab.com/tech/on-device-ai-08-application-patterns">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 11 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>performance</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[7장: 메모리 프레임워크 비교와 선택]]></title>
      <link>https://archive.kreathlab.com/tech/agent-memory-07-frameworks</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/agent-memory-07-frameworks</guid>
      <description><![CDATA[Mem0, Zep, Letta, LangChain/LangGraph의 메모리 시스템을 상세 비교하고, 프로젝트 요구사항에 맞는 프레임워크 선택 의사결정 트리를 제시합니다.]]></description>
      <content:encoded><![CDATA[<p>Mem0, Zep, Letta, LangChain/LangGraph의 메모리 시스템을 상세 비교하고, 프로젝트 요구사항에 맞는 프레임워크 선택 의사결정 트리를 제시합니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: agent-memory</p><p><a href="https://archive.kreathlab.com/tech/agent-memory-07-frameworks">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 10 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>llm</category>
      <category>ai-agent</category>
    </item>
    <item>
      <title><![CDATA[6장: 회귀 테스트 자동화]]></title>
      <link>https://archive.kreathlab.com/tech/agent-testing-06-regression</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/agent-testing-06-regression</guid>
      <description><![CDATA[평가를 회귀 테스트로 졸업시키는 패턴, Golden Dataset 관리, 롤링 성공률 모니터링, 베이스라인 관리와 변경 영향 분석을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>평가를 회귀 테스트로 졸업시키는 패턴, Golden Dataset 관리, 롤링 성공률 모니터링, 베이스라인 관리와 변경 영향 분석을 다룹니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: agent-testing</p><p><a href="https://archive.kreathlab.com/tech/agent-testing-06-regression">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 10 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>testing</category>
      <category>ai</category>
      <category>evaluation</category>
      <category>quality-assurance</category>
    </item>
    <item>
      <title><![CDATA[5장: Inspect AI - 에이전트 수준 평가]]></title>
      <link>https://archive.kreathlab.com/tech/ai-eval-harness-05-inspect-ai</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-eval-harness-05-inspect-ai</guid>
      <description><![CDATA[UK AISI의 Inspect AI를 분석합니다. 에이전트 벤치마크 GAIA, SWE-Bench, Cybench의 실행, 샌드박싱 환경, 태스크/솔버/스코러 아키텍처, 멀티에이전트 평가까지 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>UK AISI의 Inspect AI를 분석합니다. 에이전트 벤치마크 GAIA, SWE-Bench, Cybench의 실행, 샌드박싱 환경, 태스크/솔버/스코러 아키텍처, 멀티에이전트 평가까지 다룹니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: ai-eval-harness</p><p><a href="https://archive.kreathlab.com/tech/ai-eval-harness-05-inspect-ai">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 10 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>evaluation</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[8장: AI 규제와 컴플라이언스]]></title>
      <link>https://archive.kreathlab.com/tech/ai-security-08-regulation</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-security-08-regulation</guid>
      <description><![CDATA[EU AI Act를 중심으로 글로벌 AI 규제의 핵심 요구사항, 위험 분류 체계, 기술적 컴플라이언스 전략, 그리고 책임 있는 AI 개발 프레임워크를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>EU AI Act를 중심으로 글로벌 AI 규제의 핵심 요구사항, 위험 분류 체계, 기술적 컴플라이언스 전략, 그리고 책임 있는 AI 개발 프레임워크를 다룹니다.</p><p><strong>AI / ML</strong> / 13분 읽기 / 시리즈: ai-security</p><p><a href="https://archive.kreathlab.com/tech/ai-security-08-regulation">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 10 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>testing</category>
      <category>security</category>
    </item>
    <item>
      <title><![CDATA[4장: E2E 테스트 -- AI 에이전트 기반 자동화]]></title>
      <link>https://archive.kreathlab.com/tech/ai-testing-04-e2e</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-testing-04-e2e</guid>
      <description><![CDATA[자연어를 E2E 테스트로 변환하는 Momentic, testRigor, Functionize와 DOM 변경에 자동 적응하는 셀프 힐링 기능, Playwright와 AI를 결합한 실전 E2E 테스트 자동화를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>자연어를 E2E 테스트로 변환하는 Momentic, testRigor, Functionize와 DOM 변경에 자동 적응하는 셀프 힐링 기능, Playwright와 AI를 결합한 실전 E2E 테스트 자동화를 다룹니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: ai-testing</p><p><a href="https://archive.kreathlab.com/tech/ai-testing-04-e2e">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 10 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>testing</category>
      <category>automation</category>
      <category>quality-assurance</category>
      <category>ai</category>
    </item>
    <item>
      <title><![CDATA[2장: 리포지토리 인텔리전스]]></title>
      <link>https://archive.kreathlab.com/tech/context-engineering-02-repository-intelligence</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/context-engineering-02-repository-intelligence</guid>
      <description><![CDATA[AI 코딩 도구들이 코드베이스를 이해하는 방법을 비교합니다. 시맨틱 인덱싱, 코드맵, 실시간 지식 그래프, 에이전트 탐색 전략을 심층 분석합니다.]]></description>
      <content:encoded><![CDATA[<p>AI 코딩 도구들이 코드베이스를 이해하는 방법을 비교합니다. 시맨틱 인덱싱, 코드맵, 실시간 지식 그래프, 에이전트 탐색 전략을 심층 분석합니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: context-engineering</p><p><a href="https://archive.kreathlab.com/tech/context-engineering-02-repository-intelligence">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 10 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>ai-agent</category>
      <category>claude-code</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[5장: 에러 복구와 재시도 전략]]></title>
      <link>https://archive.kreathlab.com/tech/agentic-workflow-05-error-recovery</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/agentic-workflow-05-error-recovery</guid>
      <description><![CDATA[Agentic Workflow의 에러 분류 체계, 지수 백오프, 서킷 브레이커, 모델 폴백, Saga 패턴 기반 보상 트랜잭션, 데드레터 큐 등 복원력 패턴을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Agentic Workflow의 에러 분류 체계, 지수 백오프, 서킷 브레이커, 모델 폴백, Saga 패턴 기반 보상 트랜잭션, 데드레터 큐 등 복원력 패턴을 다룹니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: agentic-workflow</p><p><a href="https://archive.kreathlab.com/tech/agentic-workflow-05-error-recovery">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 09 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>workflow</category>
      <category>ai</category>
      <category>automation</category>
    </item>
    <item>
      <title><![CDATA[8장: AI 시스템의 관측 가능성]]></title>
      <link>https://archive.kreathlab.com/tech/ai-architecture-08-observability</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-architecture-08-observability</guid>
      <description><![CDATA[LLM 기반 시스템의 관측 가능성 설계 — 트레이싱, 메트릭, 로깅, 프롬프트 버전 관리, 품질 모니터링, 그리고 AI 특화 대시보드 구축을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>LLM 기반 시스템의 관측 가능성 설계 — 트레이싱, 메트릭, 로깅, 프롬프트 버전 관리, 품질 모니터링, 그리고 AI 특화 대시보드 구축을 다룹니다.</p><p><strong>아키텍처</strong> / 16분 읽기 / 시리즈: ai-architecture</p><p><a href="https://archive.kreathlab.com/tech/ai-architecture-08-observability">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 09 Mar 2026 00:00:00 GMT</pubDate>
      <category>architecture</category>
      <category>architecture</category>
      <category>llm</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[3장: 레거시 코드 이해와 문서화]]></title>
      <link>https://archive.kreathlab.com/tech/code-analysis-03-code-understanding</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/code-analysis-03-code-understanding</guid>
      <description><![CDATA[LLM을 활용하여 레거시 코드베이스를 자동으로 탐색하고 문서화하는 기법을 학습합니다. 의존성 그래프 추출, 아키텍처 다이어그램 생성, 인라인 주석 자동 생성을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>LLM을 활용하여 레거시 코드베이스를 자동으로 탐색하고 문서화하는 기법을 학습합니다. 의존성 그래프 추출, 아키텍처 다이어그램 생성, 인라인 주석 자동 생성을 다룹니다.</p><p><strong>AI / ML</strong> / 18분 읽기 / 시리즈: code-analysis</p><p><a href="https://archive.kreathlab.com/tech/code-analysis-03-code-understanding">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 09 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>code-quality</category>
      <category>ai</category>
      <category>llm</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[2장: 컨테이너 이미지 보안 기초]]></title>
      <link>https://archive.kreathlab.com/tech/container-security-02-image-security</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/container-security-02-image-security</guid>
      <description><![CDATA[최소 베이스 이미지, 멀티스테이지 빌드, 루트 없는 컨테이너 등 Dockerfile 보안 모범 사례와 불변 이미지 전략을 실습합니다.]]></description>
      <content:encoded><![CDATA[<p>최소 베이스 이미지, 멀티스테이지 빌드, 루트 없는 컨테이너 등 Dockerfile 보안 모범 사례와 불변 이미지 전략을 실습합니다.</p><p><strong>인프라</strong> / 14분 읽기 / 시리즈: container-security</p><p><a href="https://archive.kreathlab.com/tech/container-security-02-image-security">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 09 Mar 2026 00:00:00 GMT</pubDate>
      <category>infra</category>
      <category>security</category>
      <category>kubernetes</category>
      <category>devops</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[1장: 하네스 엔지니어링의 등장과 핵심 개념]]></title>
      <link>https://archive.kreathlab.com/tech/harness-engineering-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/harness-engineering-01-introduction</guid>
      <description><![CDATA[AI 에이전트에서 모델을 감싸는 모든 것, 하네스 엔지니어링의 정의와 등장 배경, 그리고 5가지 핵심 역할을 살펴봅니다.]]></description>
      <content:encoded><![CDATA[<p>AI 에이전트에서 모델을 감싸는 모든 것, 하네스 엔지니어링의 정의와 등장 배경, 그리고 5가지 핵심 역할을 살펴봅니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: harness-engineering</p><p><a href="https://archive.kreathlab.com/tech/harness-engineering-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 09 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>testing</category>
      <category>evaluation</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[7장: 엣지 하드웨어와 전용 가속기]]></title>
      <link>https://archive.kreathlab.com/tech/on-device-ai-07-edge-hardware</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/on-device-ai-07-edge-hardware</guid>
      <description><![CDATA[온디바이스 AI를 위한 하드웨어 가속기 — Apple Neural Engine, Qualcomm NPU, NVIDIA Jetson, Intel NPU의 아키텍처와 성능 특성을 비교합니다.]]></description>
      <content:encoded><![CDATA[<p>온디바이스 AI를 위한 하드웨어 가속기 — Apple Neural Engine, Qualcomm NPU, NVIDIA Jetson, Intel NPU의 아키텍처와 성능 특성을 비교합니다.</p><p><strong>AI / ML</strong> / 15분 읽기 / 시리즈: on-device-ai</p><p><a href="https://archive.kreathlab.com/tech/on-device-ai-07-edge-hardware">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 09 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>performance</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[6장: 메모리 압축과 통합]]></title>
      <link>https://archive.kreathlab.com/tech/agent-memory-06-compression</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/agent-memory-06-compression</guid>
      <description><![CDATA[에이전트 메모리의 압축 기법, 3-6배 텍스트 압축과 5-40배 도구 호출 압축, 계층적 통합과 정보 손실 최소화 전략을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>에이전트 메모리의 압축 기법, 3-6배 텍스트 압축과 5-40배 도구 호출 압축, 계층적 통합과 정보 손실 최소화 전략을 다룹니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: agent-memory</p><p><a href="https://archive.kreathlab.com/tech/agent-memory-06-compression">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 08 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>llm</category>
      <category>ai-agent</category>
    </item>
    <item>
      <title><![CDATA[5장: 비결정적 출력 평가]]></title>
      <link>https://archive.kreathlab.com/tech/agent-testing-05-nondeterministic</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/agent-testing-05-nondeterministic</guid>
      <description><![CDATA[LLM-as-Judge 패턴으로 에이전트의 비결정적 출력을 평가하는 방법, 품질 차원별 점수 산출, 임계값 설정, pass@k 전략을 상세히 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>LLM-as-Judge 패턴으로 에이전트의 비결정적 출력을 평가하는 방법, 품질 차원별 점수 산출, 임계값 설정, pass@k 전략을 상세히 다룹니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: agent-testing</p><p><a href="https://archive.kreathlab.com/tech/agent-testing-05-nondeterministic">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 08 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>testing</category>
      <category>ai</category>
      <category>evaluation</category>
      <category>quality-assurance</category>
    </item>
    <item>
      <title><![CDATA[4장: HELM - 종합적 모델 평가 프레임워크]]></title>
      <link>https://archive.kreathlab.com/tech/ai-eval-harness-04-helm</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-eval-harness-04-helm</guid>
      <description><![CDATA[Stanford CRFM의 HELM을 분석합니다. 7가지 메트릭 차원, 16가지 핵심 시나리오, HELM Lite와 MedHELM 변형, 실행 방법과 결과 분석까지 종합적 평가 접근법을 탐구합니다.]]></description>
      <content:encoded><![CDATA[<p>Stanford CRFM의 HELM을 분석합니다. 7가지 메트릭 차원, 16가지 핵심 시나리오, HELM Lite와 MedHELM 변형, 실행 방법과 결과 분석까지 종합적 평가 접근법을 탐구합니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: ai-eval-harness</p><p><a href="https://archive.kreathlab.com/tech/ai-eval-harness-04-helm">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 08 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>evaluation</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[7장: 레드티밍과 보안 테스트 자동화]]></title>
      <link>https://archive.kreathlab.com/tech/ai-security-07-red-teaming</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-security-07-red-teaming</guid>
      <description><![CDATA[AI 시스템의 레드티밍 방법론, 자동화된 보안 테스트, 프롬프트 인젝션 퍼징, 그리고 지속적 보안 검증 파이프라인 구축을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>AI 시스템의 레드티밍 방법론, 자동화된 보안 테스트, 프롬프트 인젝션 퍼징, 그리고 지속적 보안 검증 파이프라인 구축을 다룹니다.</p><p><strong>AI / ML</strong> / 14분 읽기 / 시리즈: ai-security</p><p><a href="https://archive.kreathlab.com/tech/ai-security-07-red-teaming">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 08 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>testing</category>
      <category>security</category>
    </item>
    <item>
      <title><![CDATA[3장: 통합 테스트와 API 테스트 자동화]]></title>
      <link>https://archive.kreathlab.com/tech/ai-testing-03-integration-test</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-testing-03-integration-test</guid>
      <description><![CDATA[API 스키마 기반 테스트 자동 생성, 계약 테스트(Contract Testing), testcontainers와 AI를 결합한 데이터베이스 통합 테스트, 그리고 CI 파이프라인 통합 방법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>API 스키마 기반 테스트 자동 생성, 계약 테스트(Contract Testing), testcontainers와 AI를 결합한 데이터베이스 통합 테스트, 그리고 CI 파이프라인 통합 방법을 다룹니다.</p><p><strong>AI / ML</strong> / 18분 읽기 / 시리즈: ai-testing</p><p><a href="https://archive.kreathlab.com/tech/ai-testing-03-integration-test">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 08 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>testing</category>
      <category>automation</category>
      <category>quality-assurance</category>
      <category>ai</category>
    </item>
    <item>
      <title><![CDATA[1장: 프롬프트 엔지니어링에서 컨텍스트 엔지니어링으로]]></title>
      <link>https://archive.kreathlab.com/tech/context-engineering-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/context-engineering-01-introduction</guid>
      <description><![CDATA[프롬프트 중심 사고에서 컨텍스트 설계로의 패러다임 전환을 살펴보고, 컨텍스트 엔지니어링의 정의와 5가지 핵심 전략을 개괄합니다.]]></description>
      <content:encoded><![CDATA[<p>프롬프트 중심 사고에서 컨텍스트 설계로의 패러다임 전환을 살펴보고, 컨텍스트 엔지니어링의 정의와 5가지 핵심 전략을 개괄합니다.</p><p><strong>AI / ML</strong> / 18분 읽기 / 시리즈: context-engineering</p><p><a href="https://archive.kreathlab.com/tech/context-engineering-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 08 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>ai-agent</category>
      <category>claude-code</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[4장: 장기 실행 워크플로우 관리]]></title>
      <link>https://archive.kreathlab.com/tech/agentic-workflow-04-long-running</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/agentic-workflow-04-long-running</guid>
      <description><![CDATA[시간/일 단위 워크플로우의 듀러블 실행, 체크포인팅, 일시 정지와 재개, 상태 직렬화, 타임아웃 관리, 분산 실행 전략을 정리합니다.]]></description>
      <content:encoded><![CDATA[<p>시간/일 단위 워크플로우의 듀러블 실행, 체크포인팅, 일시 정지와 재개, 상태 직렬화, 타임아웃 관리, 분산 실행 전략을 정리합니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: agentic-workflow</p><p><a href="https://archive.kreathlab.com/tech/agentic-workflow-04-long-running">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 07 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>workflow</category>
      <category>ai</category>
      <category>automation</category>
    </item>
    <item>
      <title><![CDATA[7장: 장애 대응과 회복 탄력성]]></title>
      <link>https://archive.kreathlab.com/tech/ai-architecture-07-fault-tolerance</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-architecture-07-fault-tolerance</guid>
      <description><![CDATA[AI 시스템의 장애 시나리오와 회복 탄력성 패턴 — 서킷 브레이커, 폴백, 재시도, 타임아웃, 모델 장애 조치, 그리고 그레이스풀 디그레이데이션을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>AI 시스템의 장애 시나리오와 회복 탄력성 패턴 — 서킷 브레이커, 폴백, 재시도, 타임아웃, 모델 장애 조치, 그리고 그레이스풀 디그레이데이션을 다룹니다.</p><p><strong>아키텍처</strong> / 16분 읽기 / 시리즈: ai-architecture</p><p><a href="https://archive.kreathlab.com/tech/ai-architecture-07-fault-tolerance">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 07 Mar 2026 00:00:00 GMT</pubDate>
      <category>architecture</category>
      <category>architecture</category>
      <category>llm</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[2장: AST와 LLM 하이브리드 분석]]></title>
      <link>https://archive.kreathlab.com/tech/code-analysis-02-ast-hybrid</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/code-analysis-02-ast-hybrid</guid>
      <description><![CDATA[AST 기반 정적 분석과 LLM의 의미 분석을 결합하는 하이브리드 접근법을 학습합니다. cAST 청킹, 순환 복잡도, 결합도/응집도 메트릭을 Python과 TypeScript로 실습합니다.]]></description>
      <content:encoded><![CDATA[<p>AST 기반 정적 분석과 LLM의 의미 분석을 결합하는 하이브리드 접근법을 학습합니다. cAST 청킹, 순환 복잡도, 결합도/응집도 메트릭을 Python과 TypeScript로 실습합니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: code-analysis</p><p><a href="https://archive.kreathlab.com/tech/code-analysis-02-ast-hybrid">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 07 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>code-quality</category>
      <category>ai</category>
      <category>llm</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[1장: 컨테이너 보안의 필요성과 위협 모델]]></title>
      <link>https://archive.kreathlab.com/tech/container-security-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/container-security-01-introduction</guid>
      <description><![CDATA[컨테이너 환경에서 마주하는 보안 위협과 공격 벡터를 분석하고, 방어 심층 전략과 OWASP 쿠버네티스 보안 체크리스트를 기반으로 한 보안 로드맵을 소개합니다.]]></description>
      <content:encoded><![CDATA[<p>컨테이너 환경에서 마주하는 보안 위협과 공격 벡터를 분석하고, 방어 심층 전략과 OWASP 쿠버네티스 보안 체크리스트를 기반으로 한 보안 로드맵을 소개합니다.</p><p><strong>인프라</strong> / 16분 읽기 / 시리즈: container-security</p><p><a href="https://archive.kreathlab.com/tech/container-security-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 07 Mar 2026 00:00:00 GMT</pubDate>
      <category>infra</category>
      <category>security</category>
      <category>kubernetes</category>
      <category>devops</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[6장: 모바일 디바이스에서의 AI 추론]]></title>
      <link>https://archive.kreathlab.com/tech/on-device-ai-06-mobile-inference</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/on-device-ai-06-mobile-inference</guid>
      <description><![CDATA[iOS와 Android에서의 LLM 추론 기법 — Apple MLX, Core ML, MediaPipe, Qualcomm AI Engine, 그리고 모바일 AI 앱 개발 패턴을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>iOS와 Android에서의 LLM 추론 기법 — Apple MLX, Core ML, MediaPipe, Qualcomm AI Engine, 그리고 모바일 AI 앱 개발 패턴을 다룹니다.</p><p><strong>AI / ML</strong> / 12분 읽기 / 시리즈: on-device-ai</p><p><a href="https://archive.kreathlab.com/tech/on-device-ai-06-mobile-inference">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 07 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>performance</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[11장: v3에서 v4로 마이그레이션]]></title>
      <link>https://archive.kreathlab.com/tech/tailwind-v4-11-migration</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/tailwind-v4-11-migration</guid>
      <description><![CDATA[Tailwind CSS v3에서 v4로의 마이그레이션 가이드 — 주요 변경사항, 자동 마이그레이션 도구, 단계별 전환 전략, 그리고 흔한 문제 해결법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Tailwind CSS v3에서 v4로의 마이그레이션 가이드 — 주요 변경사항, 자동 마이그레이션 도구, 단계별 전환 전략, 그리고 흔한 문제 해결법을 다룹니다.</p><p><strong>웹 개발</strong> / 13분 읽기 / 시리즈: tailwind-v4</p><p><a href="https://archive.kreathlab.com/tech/tailwind-v4-11-migration">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 07 Mar 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>frontend</category>
    </item>
    <item>
      <title><![CDATA[5장: 지식 그래프 기반 메모리 — Zep 아키텍처]]></title>
      <link>https://archive.kreathlab.com/tech/agent-memory-05-knowledge-graph</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/agent-memory-05-knowledge-graph</guid>
      <description><![CDATA[Zep의 시간 인식 동적 지식 그래프를 중심으로, 엔티티 추출, 관계 생성, 시간적 추론 등 구조화된 메모리의 설계와 장점을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Zep의 시간 인식 동적 지식 그래프를 중심으로, 엔티티 추출, 관계 생성, 시간적 추론 등 구조화된 메모리의 설계와 장점을 다룹니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: agent-memory</p><p><a href="https://archive.kreathlab.com/tech/agent-memory-05-knowledge-graph">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 06 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>llm</category>
      <category>ai-agent</category>
    </item>
    <item>
      <title><![CDATA[4장: 엔드투엔드 시나리오 테스트]]></title>
      <link>https://archive.kreathlab.com/tech/agent-testing-04-e2e</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/agent-testing-04-e2e</guid>
      <description><![CDATA[사용자 시뮬레이션 기반 멀티턴 대화 테스트, 워크플로우 완료 검증, 반복 호출 및 모순적 계획 감지 등 E2E 시나리오 테스트의 전체 방법론을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>사용자 시뮬레이션 기반 멀티턴 대화 테스트, 워크플로우 완료 검증, 반복 호출 및 모순적 계획 감지 등 E2E 시나리오 테스트의 전체 방법론을 다룹니다.</p><p><strong>AI / ML</strong> / 18분 읽기 / 시리즈: agent-testing</p><p><a href="https://archive.kreathlab.com/tech/agent-testing-04-e2e">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 06 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>testing</category>
      <category>ai</category>
      <category>evaluation</category>
      <category>quality-assurance</category>
    </item>
    <item>
      <title><![CDATA[3장: lm-evaluation-harness 심층 분석]]></title>
      <link>https://archive.kreathlab.com/tech/ai-eval-harness-03-lm-eval-harness</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-eval-harness-03-lm-eval-harness</guid>
      <description><![CDATA[EleutherAI의 lm-evaluation-harness를 심층 분석합니다. 200개 이상의 태스크, 25개 이상의 모델 백엔드, HuggingFace 리더보드 백엔드로서의 역할, 설치부터 커스텀 태스크 작성까지 실전 가이드를 제공합니다.]]></description>
      <content:encoded><![CDATA[<p>EleutherAI의 lm-evaluation-harness를 심층 분석합니다. 200개 이상의 태스크, 25개 이상의 모델 백엔드, HuggingFace 리더보드 백엔드로서의 역할, 설치부터 커스텀 태스크 작성까지 실전 가이드를 제공합니다.</p><p><strong>AI / ML</strong> / 14분 읽기 / 시리즈: ai-eval-harness</p><p><a href="https://archive.kreathlab.com/tech/ai-eval-harness-03-lm-eval-harness">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 06 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>evaluation</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[6장: LLM 애플리케이션의 인증과 권한 관리]]></title>
      <link>https://archive.kreathlab.com/tech/ai-security-06-auth-permissions</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-security-06-auth-permissions</guid>
      <description><![CDATA[LLM 기반 시스템의 인증 아키텍처, 에이전트 도구 접근 제어, 최소 권한 원칙, API 키 관리, 그리고 Human-in-the-Loop 패턴을 실전 중심으로 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>LLM 기반 시스템의 인증 아키텍처, 에이전트 도구 접근 제어, 최소 권한 원칙, API 키 관리, 그리고 Human-in-the-Loop 패턴을 실전 중심으로 다룹니다.</p><p><strong>AI / ML</strong> / 9분 읽기 / 시리즈: ai-security</p><p><a href="https://archive.kreathlab.com/tech/ai-security-06-auth-permissions">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 06 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>testing</category>
      <category>security</category>
    </item>
    <item>
      <title><![CDATA[2장: AI 기반 단위 테스트 자동 생성]]></title>
      <link>https://archive.kreathlab.com/tech/ai-testing-02-unit-test</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-testing-02-unit-test</guid>
      <description><![CDATA[LLM 기반 단위 테스트 자동 생성의 원리와 실전 활용법을 다룹니다. Diffblue, Codium/Qodo 도구를 활용한 pytest/Jest 테스트 생성 실습과 생성된 테스트의 품질 검증 방법을 안내합니다.]]></description>
      <content:encoded><![CDATA[<p>LLM 기반 단위 테스트 자동 생성의 원리와 실전 활용법을 다룹니다. Diffblue, Codium/Qodo 도구를 활용한 pytest/Jest 테스트 생성 실습과 생성된 테스트의 품질 검증 방법을 안내합니다.</p><p><strong>AI / ML</strong> / 19분 읽기 / 시리즈: ai-testing</p><p><a href="https://archive.kreathlab.com/tech/ai-testing-02-unit-test">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 06 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>testing</category>
      <category>automation</category>
      <category>quality-assurance</category>
      <category>ai</category>
    </item>
    <item>
      <title><![CDATA[11장: 운영, 모니터링, 스케일링 전략]]></title>
      <link>https://archive.kreathlab.com/tech/vector-database-11-operations</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/vector-database-11-operations</guid>
      <description><![CDATA[벡터 데이터베이스의 수평/수직 스케일링, 샤딩, 레플리카, 백업 전략, 모니터링 메트릭, 비용 최적화, 솔루션 선택 의사결정 트리, 마이그레이션 가이드를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>벡터 데이터베이스의 수평/수직 스케일링, 샤딩, 레플리카, 백업 전략, 모니터링 메트릭, 비용 최적화, 솔루션 선택 의사결정 트리, 마이그레이션 가이드를 다룹니다.</p><p><strong>AI / ML</strong> / 20분 읽기 / 시리즈: vector-database</p><p><a href="https://archive.kreathlab.com/tech/vector-database-11-operations">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 06 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>vector-database</category>
      <category>ai</category>
      <category>embedding</category>
      <category>search</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[3장: Human-in-the-Loop 설계]]></title>
      <link>https://archive.kreathlab.com/tech/agentic-workflow-03-hitl</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/agentic-workflow-03-hitl</guid>
      <description><![CDATA[Agentic Workflow의 핵심 안전장치인 HITL과 Human-on-the-Loop 패턴, 승인 게이트, 에스컬레이션 정책, 신뢰도 기반 라우팅, 점진적 자율성 확대 전략을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Agentic Workflow의 핵심 안전장치인 HITL과 Human-on-the-Loop 패턴, 승인 게이트, 에스컬레이션 정책, 신뢰도 기반 라우팅, 점진적 자율성 확대 전략을 다룹니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: agentic-workflow</p><p><a href="https://archive.kreathlab.com/tech/agentic-workflow-03-hitl">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 05 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>workflow</category>
      <category>ai</category>
      <category>automation</category>
    </item>
    <item>
      <title><![CDATA[6장: 비용 관리와 최적화 아키텍처]]></title>
      <link>https://archive.kreathlab.com/tech/ai-architecture-06-cost-management</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-architecture-06-cost-management</guid>
      <description><![CDATA[LLM API 비용을 제어하는 아키텍처 전략 — 토큰 예산 시스템, 모델 라우팅, 캐싱 경제학, 비용 모니터링, 그리고 비용 효율적 시스템 설계를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>LLM API 비용을 제어하는 아키텍처 전략 — 토큰 예산 시스템, 모델 라우팅, 캐싱 경제학, 비용 모니터링, 그리고 비용 효율적 시스템 설계를 다룹니다.</p><p><strong>아키텍처</strong> / 16분 읽기 / 시리즈: ai-architecture</p><p><a href="https://archive.kreathlab.com/tech/ai-architecture-06-cost-management">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 05 Mar 2026 00:00:00 GMT</pubDate>
      <category>architecture</category>
      <category>architecture</category>
      <category>llm</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[10장: 실전 프로젝트 - 처음부터 끝까지 Claude Code로 개발하기]]></title>
      <link>https://archive.kreathlab.com/tech/claude-code-guide-10-real-project</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/claude-code-guide-10-real-project</guid>
      <description><![CDATA[REST API 서버 프로젝트를 Claude Code만으로 처음부터 끝까지 개발하며, 앞서 배운 모든 기능을 실전에 적용합니다.]]></description>
      <content:encoded><![CDATA[<p>REST API 서버 프로젝트를 Claude Code만으로 처음부터 끝까지 개발하며, 앞서 배운 모든 기능을 실전에 적용합니다.</p><p><strong>개발 도구</strong> / 15분 읽기 / 시리즈: claude-code-guide</p><p><a href="https://archive.kreathlab.com/tech/claude-code-guide-10-real-project">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 05 Mar 2026 00:00:00 GMT</pubDate>
      <category>devtools</category>
      <category>ai-agent</category>
      <category>anthropic</category>
      <category>mcp</category>
      <category>automation</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[1장: LLM 기반 코드 분석의 등장과 가능성]]></title>
      <link>https://archive.kreathlab.com/tech/code-analysis-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/code-analysis-01-introduction</guid>
      <description><![CDATA[전통 정적 분석의 한계를 넘어 LLM이 코드를 이해하고 분석하는 새로운 패러다임을 소개합니다. 2026년 도구 생태계와 자동화 파이프라인의 가능성을 탐구합니다.]]></description>
      <content:encoded><![CDATA[<p>전통 정적 분석의 한계를 넘어 LLM이 코드를 이해하고 분석하는 새로운 패러다임을 소개합니다. 2026년 도구 생태계와 자동화 파이프라인의 가능성을 탐구합니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: code-analysis</p><p><a href="https://archive.kreathlab.com/tech/code-analysis-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 05 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>code-quality</category>
      <category>ai</category>
      <category>llm</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[5장: WebLLM과 브라우저에서의 LLM 추론]]></title>
      <link>https://archive.kreathlab.com/tech/on-device-ai-05-web-inference</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/on-device-ai-05-web-inference</guid>
      <description><![CDATA[WebGPU를 활용한 브라우저 내 LLM 추론의 원리, WebLLM과 MLC LLM의 아키텍처, 실전 구현, 그리고 브라우저 AI의 가능성과 한계를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>WebGPU를 활용한 브라우저 내 LLM 추론의 원리, WebLLM과 MLC LLM의 아키텍처, 실전 구현, 그리고 브라우저 AI의 가능성과 한계를 다룹니다.</p><p><strong>AI / ML</strong> / 12분 읽기 / 시리즈: on-device-ai</p><p><a href="https://archive.kreathlab.com/tech/on-device-ai-05-web-inference">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 05 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>performance</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[10장: 디자인 시스템 구축 실전]]></title>
      <link>https://archive.kreathlab.com/tech/tailwind-v4-10-design-system</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/tailwind-v4-10-design-system</guid>
      <description><![CDATA[Tailwind CSS v4로 프로덕션 디자인 시스템을 구축합니다. 토큰 계층 설계, 컴포넌트 라이브러리, 테마 시스템, 접근성, 팀 협업 전략을 실전 중심으로 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Tailwind CSS v4로 프로덕션 디자인 시스템을 구축합니다. 토큰 계층 설계, 컴포넌트 라이브러리, 테마 시스템, 접근성, 팀 협업 전략을 실전 중심으로 다룹니다.</p><p><strong>웹 개발</strong> / 10분 읽기 / 시리즈: tailwind-v4</p><p><a href="https://archive.kreathlab.com/tech/tailwind-v4-10-design-system">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 05 Mar 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>frontend</category>
    </item>
    <item>
      <title><![CDATA[4장: 에피소딕 메모리 — 경험에서 학습하기]]></title>
      <link>https://archive.kreathlab.com/tech/agent-memory-04-episodic</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/agent-memory-04-episodic</guid>
      <description><![CDATA[에이전트가 과거 상호작용을 에피소드로 기록하고, 경험 기반 의사결정과 패턴 학습에 활용하는 에피소딕 메모리 시스템을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>에이전트가 과거 상호작용을 에피소드로 기록하고, 경험 기반 의사결정과 패턴 학습에 활용하는 에피소딕 메모리 시스템을 다룹니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: agent-memory</p><p><a href="https://archive.kreathlab.com/tech/agent-memory-04-episodic">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 04 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>llm</category>
      <category>ai-agent</category>
    </item>
    <item>
      <title><![CDATA[3장: 도구 호출 검증]]></title>
      <link>https://archive.kreathlab.com/tech/agent-testing-03-tool-calls</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/agent-testing-03-tool-calls</guid>
      <description><![CDATA[에이전트의 도구 호출 정확성을 이름, 파라미터, 출력의 3단계로 검증하는 방법과 모킹 전략, 도구 체인 순서 검증, 불필요한 호출 감지 기법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>에이전트의 도구 호출 정확성을 이름, 파라미터, 출력의 3단계로 검증하는 방법과 모킹 전략, 도구 체인 순서 검증, 불필요한 호출 감지 기법을 다룹니다.</p><p><strong>AI / ML</strong> / 18분 읽기 / 시리즈: agent-testing</p><p><a href="https://archive.kreathlab.com/tech/agent-testing-03-tool-calls">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 04 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>testing</category>
      <category>ai</category>
      <category>evaluation</category>
      <category>quality-assurance</category>
    </item>
    <item>
      <title><![CDATA[2장: 평가 하네스 아키텍처와 핵심 개념]]></title>
      <link>https://archive.kreathlab.com/tech/ai-eval-harness-02-architecture</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-eval-harness-02-architecture</guid>
      <description><![CDATA[평가 하네스의 내부 구조를 해부합니다. 태스크 정의 시스템, 모델 백엔드 추상화, 실행 엔진의 배칭과 병렬화, 결과 집계와 리포팅까지 설계 패턴을 코드와 함께 분석합니다.]]></description>
      <content:encoded><![CDATA[<p>평가 하네스의 내부 구조를 해부합니다. 태스크 정의 시스템, 모델 백엔드 추상화, 실행 엔진의 배칭과 병렬화, 결과 집계와 리포팅까지 설계 패턴을 코드와 함께 분석합니다.</p><p><strong>AI / ML</strong> / 18분 읽기 / 시리즈: ai-eval-harness</p><p><a href="https://archive.kreathlab.com/tech/ai-eval-harness-02-architecture">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 04 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>evaluation</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[5장: 콘텐츠 안전성과 유해 출력 방지]]></title>
      <link>https://archive.kreathlab.com/tech/ai-security-05-content-safety</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-security-05-content-safety</guid>
      <description><![CDATA[LLM의 유해 콘텐츠 생성 방지, 편향 완화, 환각 탐지, 그리고 Constitutional AI와 RLHF의 원리를 다루며 안전한 AI 출력을 위한 다층 전략을 설계합니다.]]></description>
      <content:encoded><![CDATA[<p>LLM의 유해 콘텐츠 생성 방지, 편향 완화, 환각 탐지, 그리고 Constitutional AI와 RLHF의 원리를 다루며 안전한 AI 출력을 위한 다층 전략을 설계합니다.</p><p><strong>AI / ML</strong> / 10분 읽기 / 시리즈: ai-security</p><p><a href="https://archive.kreathlab.com/tech/ai-security-05-content-safety">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 04 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>testing</category>
      <category>security</category>
    </item>
    <item>
      <title><![CDATA[1장: AI 기반 테스트 자동화의 진화와 현재]]></title>
      <link>https://archive.kreathlab.com/tech/ai-testing-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-testing-01-introduction</guid>
      <description><![CDATA[전통적인 테스트 자동화에서 AI 기반 테스트로의 전환을 살펴봅니다. Agentic QA의 등장, 2026년 도구 생태계, 그리고 70% 이상 기업이 도입한 AI 테스트의 현황과 30-45% 효율 개선 사례를 분석합니다.]]></description>
      <content:encoded><![CDATA[<p>전통적인 테스트 자동화에서 AI 기반 테스트로의 전환을 살펴봅니다. Agentic QA의 등장, 2026년 도구 생태계, 그리고 70% 이상 기업이 도입한 AI 테스트의 현황과 30-45% 효율 개선 사례를 분석합니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: ai-testing</p><p><a href="https://archive.kreathlab.com/tech/ai-testing-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 04 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>testing</category>
      <category>automation</category>
      <category>quality-assurance</category>
      <category>ai</category>
    </item>
    <item>
      <title><![CDATA[10장: 메타데이터 필터링과 고급 쿼리]]></title>
      <link>https://archive.kreathlab.com/tech/vector-database-10-metadata-filtering</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/vector-database-10-metadata-filtering</guid>
      <description><![CDATA[사전 필터링과 사후 필터링의 차이, 필터 인덱스 설계, 복합 필터 조건, 지오 필터, 멀티테넌시 필터 패턴, 성능 최적화 전략을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>사전 필터링과 사후 필터링의 차이, 필터 인덱스 설계, 복합 필터 조건, 지오 필터, 멀티테넌시 필터 패턴, 성능 최적화 전략을 다룹니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: vector-database</p><p><a href="https://archive.kreathlab.com/tech/vector-database-10-metadata-filtering">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 04 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>vector-database</category>
      <category>ai</category>
      <category>embedding</category>
      <category>search</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[2장: 에이전트 기반 비즈니스 프로세스 설계]]></title>
      <link>https://archive.kreathlab.com/tech/agentic-workflow-02-design</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/agentic-workflow-02-design</guid>
      <description><![CDATA[Agentic Workflow를 위한 프로세스 분석, 태스크 분해, 에이전트 역할 정의, 워크플로우 DAG 설계 방법론을 체계적으로 정리합니다.]]></description>
      <content:encoded><![CDATA[<p>Agentic Workflow를 위한 프로세스 분석, 태스크 분해, 에이전트 역할 정의, 워크플로우 DAG 설계 방법론을 체계적으로 정리합니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: agentic-workflow</p><p><a href="https://archive.kreathlab.com/tech/agentic-workflow-02-design">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 03 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>workflow</category>
      <category>ai</category>
      <category>automation</category>
    </item>
    <item>
      <title><![CDATA[5장: AI 시스템의 캐싱 전략]]></title>
      <link>https://archive.kreathlab.com/tech/ai-architecture-05-caching</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-architecture-05-caching</guid>
      <description><![CDATA[LLM 추론 비용과 지연을 줄이는 다층 캐싱 전략 — 의미론적 캐시, 프롬프트 캐시, KV 캐시, 임베딩 캐시, 그리고 캐시 무효화 전략을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>LLM 추론 비용과 지연을 줄이는 다층 캐싱 전략 — 의미론적 캐시, 프롬프트 캐시, KV 캐시, 임베딩 캐시, 그리고 캐시 무효화 전략을 다룹니다.</p><p><strong>아키텍처</strong> / 11분 읽기 / 시리즈: ai-architecture</p><p><a href="https://archive.kreathlab.com/tech/ai-architecture-05-caching">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 03 Mar 2026 00:00:00 GMT</pubDate>
      <category>architecture</category>
      <category>architecture</category>
      <category>llm</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[9장: Git 워크플로우와 CI/CD 통합]]></title>
      <link>https://archive.kreathlab.com/tech/claude-code-guide-09-git-cicd</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/claude-code-guide-09-git-cicd</guid>
      <description><![CDATA[Claude Code를 Git 워크플로우에 통합하고, CI/CD 파이프라인에서 자동화된 코드 리뷰와 수정을 구성하는 방법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Claude Code를 Git 워크플로우에 통합하고, CI/CD 파이프라인에서 자동화된 코드 리뷰와 수정을 구성하는 방법을 다룹니다.</p><p><strong>개발 도구</strong> / 14분 읽기 / 시리즈: claude-code-guide</p><p><a href="https://archive.kreathlab.com/tech/claude-code-guide-09-git-cicd">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 03 Mar 2026 00:00:00 GMT</pubDate>
      <category>devtools</category>
      <category>ai-agent</category>
      <category>anthropic</category>
      <category>mcp</category>
      <category>automation</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[4장: AWQ, GPTQ, 양자화 기법 비교]]></title>
      <link>https://archive.kreathlab.com/tech/on-device-ai-04-awq-gptq</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/on-device-ai-04-awq-gptq</guid>
      <description><![CDATA[GPU 추론에 최적화된 AWQ와 GPTQ 양자화 기법의 원리, 차이점, 그리고 GGUF와의 비교를 통해 상황별 최적 양자화 전략을 제시합니다.]]></description>
      <content:encoded><![CDATA[<p>GPU 추론에 최적화된 AWQ와 GPTQ 양자화 기법의 원리, 차이점, 그리고 GGUF와의 비교를 통해 상황별 최적 양자화 전략을 제시합니다.</p><p><strong>AI / ML</strong> / 12분 읽기 / 시리즈: on-device-ai</p><p><a href="https://archive.kreathlab.com/tech/on-device-ai-04-awq-gptq">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 03 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>performance</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[11장: 실전 프로젝트 -- Rust 백엔드 API 구축]]></title>
      <link>https://archive.kreathlab.com/tech/rust-backend-11-project</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/rust-backend-11-project</guid>
      <description><![CDATA[Axum, SQLx, Tokio를 조합한 프로덕션 수준의 REST API를 처음부터 구축합니다. JWT 인증, CRUD, 미들웨어, 테스트, Docker 배포까지 총정리합니다.]]></description>
      <content:encoded><![CDATA[<p>Axum, SQLx, Tokio를 조합한 프로덕션 수준의 REST API를 처음부터 구축합니다. JWT 인증, CRUD, 미들웨어, 테스트, Docker 배포까지 총정리합니다.</p><p><strong>프로그래밍</strong> / 16분 읽기 / 시리즈: rust-backend</p><p><a href="https://archive.kreathlab.com/tech/rust-backend-11-project">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 03 Mar 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>rust</category>
    </item>
    <item>
      <title><![CDATA[9장: 커스텀 변형과 플러그인 시스템]]></title>
      <link>https://archive.kreathlab.com/tech/tailwind-v4-09-plugins-variants</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/tailwind-v4-09-plugins-variants</guid>
      <description><![CDATA[Tailwind CSS v4의 @custom-variant, @plugin 디렉티브, CSS 기반 확장 시스템, 그리고 기존 JS 플러그인에서의 마이그레이션을 실전 중심으로 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Tailwind CSS v4의 @custom-variant, @plugin 디렉티브, CSS 기반 확장 시스템, 그리고 기존 JS 플러그인에서의 마이그레이션을 실전 중심으로 다룹니다.</p><p><strong>웹 개발</strong> / 9분 읽기 / 시리즈: tailwind-v4</p><p><a href="https://archive.kreathlab.com/tech/tailwind-v4-09-plugins-variants">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 03 Mar 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>frontend</category>
    </item>
    <item>
      <title><![CDATA[3장: 장기 메모리 — 벡터 기반 의미 검색]]></title>
      <link>https://archive.kreathlab.com/tech/agent-memory-03-long-term</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/agent-memory-03-long-term</guid>
      <description><![CDATA[벡터 데이터베이스에 메모리를 저장하고 임베딩 기반으로 검색하는 장기 메모리 시스템의 설계와 구현 전략을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>벡터 데이터베이스에 메모리를 저장하고 임베딩 기반으로 검색하는 장기 메모리 시스템의 설계와 구현 전략을 다룹니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: agent-memory</p><p><a href="https://archive.kreathlab.com/tech/agent-memory-03-long-term">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 02 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>llm</category>
      <category>ai-agent</category>
    </item>
    <item>
      <title><![CDATA[2장: 에이전트 행동 테스트 프레임워크]]></title>
      <link>https://archive.kreathlab.com/tech/agent-testing-02-frameworks</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/agent-testing-02-frameworks</guid>
      <description><![CDATA[Scenario, Agentest, Inspect AI, Braintrust 등 주요 에이전트 테스트 프레임워크를 비교하고, 프로젝트 특성에 맞는 선택 기준과 환경 설정 방법을 안내합니다.]]></description>
      <content:encoded><![CDATA[<p>Scenario, Agentest, Inspect AI, Braintrust 등 주요 에이전트 테스트 프레임워크를 비교하고, 프로젝트 특성에 맞는 선택 기준과 환경 설정 방법을 안내합니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: agent-testing</p><p><a href="https://archive.kreathlab.com/tech/agent-testing-02-frameworks">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 02 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>testing</category>
      <category>ai</category>
      <category>evaluation</category>
      <category>quality-assurance</category>
    </item>
    <item>
      <title><![CDATA[1장: AI 평가의 현재와 평가 하네스의 역할]]></title>
      <link>https://archive.kreathlab.com/tech/ai-eval-harness-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-eval-harness-01-introduction</guid>
      <description><![CDATA[300개 이상의 모델과 50개 이상의 벤치마크가 공존하는 시대, AI 평가 하네스가 왜 필요한지 그 정의와 핵심 구성요소, 평가 생태계 전체 지도를 살펴봅니다.]]></description>
      <content:encoded><![CDATA[<p>300개 이상의 모델과 50개 이상의 벤치마크가 공존하는 시대, AI 평가 하네스가 왜 필요한지 그 정의와 핵심 구성요소, 평가 생태계 전체 지도를 살펴봅니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: ai-eval-harness</p><p><a href="https://archive.kreathlab.com/tech/ai-eval-harness-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 02 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>evaluation</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[4장: 가드레일 설계 — 입력/출력 필터링 시스템]]></title>
      <link>https://archive.kreathlab.com/tech/ai-security-04-guardrails</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-security-04-guardrails</guid>
      <description><![CDATA[LLM 가드레일 시스템의 설계 원리, Llama Guard, NeMo Guardrails, Guardrails AI 등 주요 도구의 비교와 활용, 그리고 커스텀 가드레일 구축을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>LLM 가드레일 시스템의 설계 원리, Llama Guard, NeMo Guardrails, Guardrails AI 등 주요 도구의 비교와 활용, 그리고 커스텀 가드레일 구축을 다룹니다.</p><p><strong>AI / ML</strong> / 12분 읽기 / 시리즈: ai-security</p><p><a href="https://archive.kreathlab.com/tech/ai-security-04-guardrails">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 02 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>testing</category>
      <category>security</category>
    </item>
    <item>
      <title><![CDATA[11장: 프로덕션 모니터링과 운영]]></title>
      <link>https://archive.kreathlab.com/tech/realtime-pipeline-11-operations</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/realtime-pipeline-11-operations</guid>
      <description><![CDATA[브로커/프로듀서/컨슈머 핵심 메트릭, Prometheus/Grafana 대시보드, 알림 설계, 용량 계획, 비용 최적화, 실전 아키텍처까지 실시간 파이프라인의 프로덕션 운영을 학습합니다.]]></description>
      <content:encoded><![CDATA[<p>브로커/프로듀서/컨슈머 핵심 메트릭, Prometheus/Grafana 대시보드, 알림 설계, 용량 계획, 비용 최적화, 실전 아키텍처까지 실시간 파이프라인의 프로덕션 운영을 학습합니다.</p><p><strong>데이터</strong> / 23분 읽기 / 시리즈: realtime-pipeline</p><p><a href="https://archive.kreathlab.com/tech/realtime-pipeline-11-operations">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 02 Mar 2026 00:00:00 GMT</pubDate>
      <category>data</category>
      <category>streaming</category>
      <category>data-engineering</category>
    </item>
    <item>
      <title><![CDATA[9장: 하이브리드 검색 구현]]></title>
      <link>https://archive.kreathlab.com/tech/vector-database-09-hybrid-search</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/vector-database-09-hybrid-search</guid>
      <description><![CDATA[시맨틱 검색과 키워드 검색을 결합하는 하이브리드 검색의 원리, BM25+벡터 퓨전 전략, Reciprocal Rank Fusion, 리랭커 통합, 프레임워크별 구현 방법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>시맨틱 검색과 키워드 검색을 결합하는 하이브리드 검색의 원리, BM25+벡터 퓨전 전략, Reciprocal Rank Fusion, 리랭커 통합, 프레임워크별 구현 방법을 다룹니다.</p><p><strong>AI / ML</strong> / 14분 읽기 / 시리즈: vector-database</p><p><a href="https://archive.kreathlab.com/tech/vector-database-09-hybrid-search">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 02 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>vector-database</category>
      <category>ai</category>
      <category>embedding</category>
      <category>search</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[1장: Agentic Workflow의 등장과 핵심 개념]]></title>
      <link>https://archive.kreathlab.com/tech/agentic-workflow-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/agentic-workflow-01-introduction</guid>
      <description><![CDATA[전통적 자동화에서 에이전틱 자동화로의 전환을 살펴보고, Agentic Workflow의 핵심 개념과 3단계 진화 모델, 구성 요소, 실제 사용 사례를 정리합니다.]]></description>
      <content:encoded><![CDATA[<p>전통적 자동화에서 에이전틱 자동화로의 전환을 살펴보고, Agentic Workflow의 핵심 개념과 3단계 진화 모델, 구성 요소, 실제 사용 사례를 정리합니다.</p><p><strong>AI / ML</strong> / 18분 읽기 / 시리즈: agentic-workflow</p><p><a href="https://archive.kreathlab.com/tech/agentic-workflow-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 01 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>workflow</category>
      <category>ai</category>
      <category>automation</category>
    </item>
    <item>
      <title><![CDATA[4장: 이벤트 드리븐 AI 파이프라인]]></title>
      <link>https://archive.kreathlab.com/tech/ai-architecture-04-event-driven</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-architecture-04-event-driven</guid>
      <description><![CDATA[이벤트 기반 아키텍처로 AI 워크로드를 처리하는 패턴 — 메시지 큐, 스트리밍 처리, 비동기 추론, 그리고 실시간 AI 파이프라인 설계를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>이벤트 기반 아키텍처로 AI 워크로드를 처리하는 패턴 — 메시지 큐, 스트리밍 처리, 비동기 추론, 그리고 실시간 AI 파이프라인 설계를 다룹니다.</p><p><strong>아키텍처</strong> / 10분 읽기 / 시리즈: ai-architecture</p><p><a href="https://archive.kreathlab.com/tech/ai-architecture-04-event-driven">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 01 Mar 2026 00:00:00 GMT</pubDate>
      <category>architecture</category>
      <category>architecture</category>
      <category>llm</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[8장: 서브에이전트와 에이전트 팀 활용하기]]></title>
      <link>https://archive.kreathlab.com/tech/claude-code-guide-08-agents</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/claude-code-guide-08-agents</guid>
      <description><![CDATA[Claude Code의 서브에이전트와 에이전트 팀 기능을 이해하고, 복잡한 작업을 병렬로 분할 처리하는 방법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Claude Code의 서브에이전트와 에이전트 팀 기능을 이해하고, 복잡한 작업을 병렬로 분할 처리하는 방법을 다룹니다.</p><p><strong>개발 도구</strong> / 16분 읽기 / 시리즈: claude-code-guide</p><p><a href="https://archive.kreathlab.com/tech/claude-code-guide-08-agents">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 01 Mar 2026 00:00:00 GMT</pubDate>
      <category>devtools</category>
      <category>ai-agent</category>
      <category>anthropic</category>
      <category>mcp</category>
      <category>automation</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[3장: GGUF와 llama.cpp 생태계]]></title>
      <link>https://archive.kreathlab.com/tech/on-device-ai-03-gguf-llamacpp</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/on-device-ai-03-gguf-llamacpp</guid>
      <description><![CDATA[온디바이스 LLM 추론의 핵심인 GGUF 형식과 llama.cpp의 아키텍처, 설치와 사용법, 성능 최적화, 그리고 주요 프론트엔드 도구를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>온디바이스 LLM 추론의 핵심인 GGUF 형식과 llama.cpp의 아키텍처, 설치와 사용법, 성능 최적화, 그리고 주요 프론트엔드 도구를 다룹니다.</p><p><strong>AI / ML</strong> / 10분 읽기 / 시리즈: on-device-ai</p><p><a href="https://archive.kreathlab.com/tech/on-device-ai-03-gguf-llamacpp">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 01 Mar 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>performance</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[10장: WebAssembly 타겟]]></title>
      <link>https://archive.kreathlab.com/tech/rust-backend-10-wasm</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/rust-backend-10-wasm</guid>
      <description><![CDATA[wasm-pack, wasm-bindgen, wasm32-wasi 타겟, Spin 서버리스, 브라우저 통합, 서버와 Wasm 간 비즈니스 로직 공유, 크기 최적화까지 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>wasm-pack, wasm-bindgen, wasm32-wasi 타겟, Spin 서버리스, 브라우저 통합, 서버와 Wasm 간 비즈니스 로직 공유, 크기 최적화까지 다룹니다.</p><p><strong>프로그래밍</strong> / 13분 읽기 / 시리즈: rust-backend</p><p><a href="https://archive.kreathlab.com/tech/rust-backend-10-wasm">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 01 Mar 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>rust</category>
    </item>
    <item>
      <title><![CDATA[8장: 다크 모드와 테마 전략]]></title>
      <link>https://archive.kreathlab.com/tech/tailwind-v4-08-dark-mode-theming</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/tailwind-v4-08-dark-mode-theming</guid>
      <description><![CDATA[Tailwind CSS v4의 다크 모드 전략, CSS 커스텀 프로퍼티 기반 런타임 테마 전환, 다중 테마 패턴, 그리고 next-themes와의 통합을 실전 중심으로 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Tailwind CSS v4의 다크 모드 전략, CSS 커스텀 프로퍼티 기반 런타임 테마 전환, 다중 테마 패턴, 그리고 next-themes와의 통합을 실전 중심으로 다룹니다.</p><p><strong>웹 개발</strong> / 10분 읽기 / 시리즈: tailwind-v4</p><p><a href="https://archive.kreathlab.com/tech/tailwind-v4-08-dark-mode-theming">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 01 Mar 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>frontend</category>
    </item>
    <item>
      <title><![CDATA[2장: 단기 메모리와 컨텍스트 윈도우 관리]]></title>
      <link>https://archive.kreathlab.com/tech/agent-memory-02-short-term</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/agent-memory-02-short-term</guid>
      <description><![CDATA[슬라이딩 윈도우, 메시지 요약, 토큰 예산 관리, 중요도 기반 정리 등 에이전트 단기 메모리의 핵심 전략을 코드 예제와 함께 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>슬라이딩 윈도우, 메시지 요약, 토큰 예산 관리, 중요도 기반 정리 등 에이전트 단기 메모리의 핵심 전략을 코드 예제와 함께 다룹니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: agent-memory</p><p><a href="https://archive.kreathlab.com/tech/agent-memory-02-short-term">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 28 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>llm</category>
      <category>ai-agent</category>
    </item>
    <item>
      <title><![CDATA[1장: AI 에이전트 테스트의 필요성과 과제]]></title>
      <link>https://archive.kreathlab.com/tech/agent-testing-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/agent-testing-01-introduction</guid>
      <description><![CDATA[결정론적 소프트웨어와 비결정적 AI 에이전트의 근본적 차이를 짚고, 전통적 QA 방법론의 한계와 에이전트 전용 3단계 테스트 모델을 소개합니다.]]></description>
      <content:encoded><![CDATA[<p>결정론적 소프트웨어와 비결정적 AI 에이전트의 근본적 차이를 짚고, 전통적 QA 방법론의 한계와 에이전트 전용 3단계 테스트 모델을 소개합니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: agent-testing</p><p><a href="https://archive.kreathlab.com/tech/agent-testing-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 28 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>testing</category>
      <category>ai</category>
      <category>evaluation</category>
      <category>quality-assurance</category>
    </item>
    <item>
      <title><![CDATA[3장: 간접 프롬프트 인젝션과 데이터 오염]]></title>
      <link>https://archive.kreathlab.com/tech/ai-security-03-indirect-injection</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-security-03-indirect-injection</guid>
      <description><![CDATA[간접 프롬프트 인젝션의 공격 벡터, RAG 오염, 이메일/웹 기반 공격, 그리고 데이터 소스 신뢰도 관리와 방어 전략을 실전 중심으로 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>간접 프롬프트 인젝션의 공격 벡터, RAG 오염, 이메일/웹 기반 공격, 그리고 데이터 소스 신뢰도 관리와 방어 전략을 실전 중심으로 다룹니다.</p><p><strong>AI / ML</strong> / 12분 읽기 / 시리즈: ai-security</p><p><a href="https://archive.kreathlab.com/tech/ai-security-03-indirect-injection">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 28 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>testing</category>
      <category>security</category>
    </item>
    <item>
      <title><![CDATA[11장: 실전 프로젝트 -- 관측 가능성 플랫폼 구축]]></title>
      <link>https://archive.kreathlab.com/tech/opentelemetry-11-project</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/opentelemetry-11-project</guid>
      <description><![CDATA[마이크로서비스 계측, AI 서비스 관측, SLO 알림을 종합하여 전체 관측 가능성 플랫폼을 구축하고, 운영 체크리스트와 비용 최적화 전략을 정리합니다.]]></description>
      <content:encoded><![CDATA[<p>마이크로서비스 계측, AI 서비스 관측, SLO 알림을 종합하여 전체 관측 가능성 플랫폼을 구축하고, 운영 체크리스트와 비용 최적화 전략을 정리합니다.</p><p><strong>인프라</strong> / 14분 읽기 / 시리즈: opentelemetry</p><p><a href="https://archive.kreathlab.com/tech/opentelemetry-11-project">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 28 Feb 2026 00:00:00 GMT</pubDate>
      <category>infra</category>
      <category>monitoring</category>
      <category>observability</category>
    </item>
    <item>
      <title><![CDATA[10장: Exactly-once 보장과 신뢰성]]></title>
      <link>https://archive.kreathlab.com/tech/realtime-pipeline-10-reliability</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/realtime-pipeline-10-reliability</guid>
      <description><![CDATA[At-least-once/At-most-once/Exactly-once 비교, Kafka 트랜잭션과 Flink 체크포인트의 조합, 멱등성 설계, DLQ, 백프레셔, 장애 복구 전략까지 프로덕션 신뢰성을 학습합니다.]]></description>
      <content:encoded><![CDATA[<p>At-least-once/At-most-once/Exactly-once 비교, Kafka 트랜잭션과 Flink 체크포인트의 조합, 멱등성 설계, DLQ, 백프레셔, 장애 복구 전략까지 프로덕션 신뢰성을 학습합니다.</p><p><strong>데이터</strong> / 19분 읽기 / 시리즈: realtime-pipeline</p><p><a href="https://archive.kreathlab.com/tech/realtime-pipeline-10-reliability">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 28 Feb 2026 00:00:00 GMT</pubDate>
      <category>data</category>
      <category>streaming</category>
      <category>data-engineering</category>
    </item>
    <item>
      <title><![CDATA[8장: Qdrant와 pgvector -- 특화 솔루션들]]></title>
      <link>https://archive.kreathlab.com/tech/vector-database-08-qdrant-pgvector</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/vector-database-08-qdrant-pgvector</guid>
      <description><![CDATA[Rust 기반 고성능 벡터 엔진 Qdrant의 페이로드 필터링, 명명된 벡터, 하이브리드 배포를 분석하고, PostgreSQL 확장 pgvector의 트랜잭션 일관성과 pgvectorscale 성능을 비교합니다.]]></description>
      <content:encoded><![CDATA[<p>Rust 기반 고성능 벡터 엔진 Qdrant의 페이로드 필터링, 명명된 벡터, 하이브리드 배포를 분석하고, PostgreSQL 확장 pgvector의 트랜잭션 일관성과 pgvectorscale 성능을 비교합니다.</p><p><strong>AI / ML</strong> / 14분 읽기 / 시리즈: vector-database</p><p><a href="https://archive.kreathlab.com/tech/vector-database-08-qdrant-pgvector">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 28 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>vector-database</category>
      <category>ai</category>
      <category>embedding</category>
      <category>search</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[3장: LLM 통합 아키텍처 패턴]]></title>
      <link>https://archive.kreathlab.com/tech/ai-architecture-03-llm-integration</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-architecture-03-llm-integration</guid>
      <description><![CDATA[LLM을 시스템에 통합하는 핵심 아키텍처 패턴 — Gateway 패턴, Router 패턴, Chain 패턴, Orchestrator 패턴, 그리고 RAG 아키텍처의 설계를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>LLM을 시스템에 통합하는 핵심 아키텍처 패턴 — Gateway 패턴, Router 패턴, Chain 패턴, Orchestrator 패턴, 그리고 RAG 아키텍처의 설계를 다룹니다.</p><p><strong>아키텍처</strong> / 11분 읽기 / 시리즈: ai-architecture</p><p><a href="https://archive.kreathlab.com/tech/ai-architecture-03-llm-integration">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 27 Feb 2026 00:00:00 GMT</pubDate>
      <category>architecture</category>
      <category>architecture</category>
      <category>llm</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[7장: MCP 서버 연동으로 외부 도구 통합하기]]></title>
      <link>https://archive.kreathlab.com/tech/claude-code-guide-07-mcp</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/claude-code-guide-07-mcp</guid>
      <description><![CDATA[Model Context Protocol의 개념을 이해하고, Claude Code에 MCP 서버를 연결하여 데이터베이스, API, 외부 서비스를 통합하는 방법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Model Context Protocol의 개념을 이해하고, Claude Code에 MCP 서버를 연결하여 데이터베이스, API, 외부 서비스를 통합하는 방법을 다룹니다.</p><p><strong>개발 도구</strong> / 13분 읽기 / 시리즈: claude-code-guide</p><p><a href="https://archive.kreathlab.com/tech/claude-code-guide-07-mcp">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 27 Feb 2026 00:00:00 GMT</pubDate>
      <category>devtools</category>
      <category>ai-agent</category>
      <category>anthropic</category>
      <category>mcp</category>
      <category>automation</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[11장: 실전 프로젝트 — 멀티모달 AI 애플리케이션 구축]]></title>
      <link>https://archive.kreathlab.com/tech/multimodal-ai-11-practical-project</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/multimodal-ai-11-practical-project</guid>
      <description><![CDATA[시리즈 전체의 기법을 종합하여 멀티모달 문서 분석 시스템을 설계하고 구현합니다. 이미지, 표, 차트를 이해하는 RAG 기반 Q&A 시스템을 구축합니다.]]></description>
      <content:encoded><![CDATA[<p>시리즈 전체의 기법을 종합하여 멀티모달 문서 분석 시스템을 설계하고 구현합니다. 이미지, 표, 차트를 이해하는 RAG 기반 Q&amp;A 시스템을 구축합니다.</p><p><strong>AI / ML</strong> / 11분 읽기 / 시리즈: multimodal-ai</p><p><a href="https://archive.kreathlab.com/tech/multimodal-ai-11-practical-project">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 27 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>multimodal</category>
      <category>embedding</category>
    </item>
    <item>
      <title><![CDATA[2장: 모델 양자화 기초 — 정밀도, 형식, 트레이드오프]]></title>
      <link>https://archive.kreathlab.com/tech/on-device-ai-02-quantization-basics</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/on-device-ai-02-quantization-basics</guid>
      <description><![CDATA[모델 양자화의 핵심 원리, 부동소수점과 정수 표현, 양자화 형식(대칭/비대칭, 채널/그룹), 품질-크기 트레이드오프를 체계적으로 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>모델 양자화의 핵심 원리, 부동소수점과 정수 표현, 양자화 형식(대칭/비대칭, 채널/그룹), 품질-크기 트레이드오프를 체계적으로 다룹니다.</p><p><strong>AI / ML</strong> / 11분 읽기 / 시리즈: on-device-ai</p><p><a href="https://archive.kreathlab.com/tech/on-device-ai-02-quantization-basics">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 27 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>performance</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[9장: CLI 도구 개발]]></title>
      <link>https://archive.kreathlab.com/tech/rust-backend-09-cli</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/rust-backend-09-cli</guid>
      <description><![CDATA[Clap v4 서브커맨드 아키텍처, config-rs 설정 관리, tracing 구조화 로깅, indicatif 진행 표시, 크로스 플랫폼 빌드와 배포까지 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Clap v4 서브커맨드 아키텍처, config-rs 설정 관리, tracing 구조화 로깅, indicatif 진행 표시, 크로스 플랫폼 빌드와 배포까지 다룹니다.</p><p><strong>프로그래밍</strong> / 12분 읽기 / 시리즈: rust-backend</p><p><a href="https://archive.kreathlab.com/tech/rust-backend-09-cli">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 27 Feb 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>rust</category>
    </item>
    <item>
      <title><![CDATA[7장: 고급 애니메이션과 트랜지션]]></title>
      <link>https://archive.kreathlab.com/tech/tailwind-v4-07-animations</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/tailwind-v4-07-animations</guid>
      <description><![CDATA[Tailwind CSS v4의 애니메이션 시스템, @keyframes 정의, @starting-style 활용, transition-behavior, 그리고 성능 최적화된 애니메이션 패턴을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Tailwind CSS v4의 애니메이션 시스템, @keyframes 정의, @starting-style 활용, transition-behavior, 그리고 성능 최적화된 애니메이션 패턴을 다룹니다.</p><p><strong>웹 개발</strong> / 11분 읽기 / 시리즈: tailwind-v4</p><p><a href="https://archive.kreathlab.com/tech/tailwind-v4-07-animations">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 27 Feb 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>frontend</category>
    </item>
    <item>
      <title><![CDATA[1장: AI 에이전트 메모리의 필요성과 핵심 개념]]></title>
      <link>https://archive.kreathlab.com/tech/agent-memory-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/agent-memory-01-introduction</guid>
      <description><![CDATA[AI 에이전트가 왜 외부 메모리를 필요로 하는지, 컨텍스트 윈도우의 한계와 3가지 메모리 유형(단기/장기/에피소딕), 2026년 메모리 프레임워크 생태계를 개괄합니다.]]></description>
      <content:encoded><![CDATA[<p>AI 에이전트가 왜 외부 메모리를 필요로 하는지, 컨텍스트 윈도우의 한계와 3가지 메모리 유형(단기/장기/에피소딕), 2026년 메모리 프레임워크 생태계를 개괄합니다.</p><p><strong>AI / ML</strong> / 14분 읽기 / 시리즈: agent-memory</p><p><a href="https://archive.kreathlab.com/tech/agent-memory-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 26 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai</category>
      <category>llm</category>
      <category>ai-agent</category>
    </item>
    <item>
      <title><![CDATA[2장: 프롬프트 인젝션 공격과 방어]]></title>
      <link>https://archive.kreathlab.com/tech/ai-security-02-prompt-injection</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-security-02-prompt-injection</guid>
      <description><![CDATA[직접 프롬프트 인젝션의 공격 기법, 탈옥 패턴, 그리고 인스트럭션 계층, 입력 검증, 구분자 전략 등 실전 방어 기법을 체계적으로 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>직접 프롬프트 인젝션의 공격 기법, 탈옥 패턴, 그리고 인스트럭션 계층, 입력 검증, 구분자 전략 등 실전 방어 기법을 체계적으로 다룹니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: ai-security</p><p><a href="https://archive.kreathlab.com/tech/ai-security-02-prompt-injection">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 26 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>testing</category>
      <category>security</category>
    </item>
    <item>
      <title><![CDATA[11장: 실전 프로젝트 -- GraphQL API 구축]]></title>
      <link>https://archive.kreathlab.com/tech/graphql-architecture-11-project</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/graphql-architecture-11-project</guid>
      <description><![CDATA[시리즈의 모든 개념을 종합하여 Apollo Server, Federation, DataLoader 기반의 프로덕션 수준 GraphQL API를 구축합니다. 스키마 설계부터 Docker 배포, 성능 테스트까지 전 과정을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>시리즈의 모든 개념을 종합하여 Apollo Server, Federation, DataLoader 기반의 프로덕션 수준 GraphQL API를 구축합니다. 스키마 설계부터 Docker 배포, 성능 테스트까지 전 과정을 다룹니다.</p><p><strong>웹 개발</strong> / 17분 읽기 / 시리즈: graphql-architecture</p><p><a href="https://archive.kreathlab.com/tech/graphql-architecture-11-project">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 26 Feb 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>graphql</category>
      <category>api-design</category>
      <category>frontend</category>
    </item>
    <item>
      <title><![CDATA[10장: SLO 기반 알림 설계]]></title>
      <link>https://archive.kreathlab.com/tech/opentelemetry-10-slo</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/opentelemetry-10-slo</guid>
      <description><![CDATA[SLI/SLO/에러 버짓의 개념을 정립하고, 번 레이트 알림(fast-burn/slow-burn), Prometheus 알림 규칙, Grafana 알림 채널을 설계합니다.]]></description>
      <content:encoded><![CDATA[<p>SLI/SLO/에러 버짓의 개념을 정립하고, 번 레이트 알림(fast-burn/slow-burn), Prometheus 알림 규칙, Grafana 알림 채널을 설계합니다.</p><p><strong>인프라</strong> / 15분 읽기 / 시리즈: opentelemetry</p><p><a href="https://archive.kreathlab.com/tech/opentelemetry-10-slo">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 26 Feb 2026 00:00:00 GMT</pubDate>
      <category>infra</category>
      <category>monitoring</category>
      <category>observability</category>
    </item>
    <item>
      <title><![CDATA[9장: 스키마 레지스트리와 데이터 계약]]></title>
      <link>https://archive.kreathlab.com/tech/realtime-pipeline-09-schema</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/realtime-pipeline-09-schema</guid>
      <description><![CDATA[스키마 진화의 필요성, Confluent Schema Registry, Avro/Protobuf/JSON Schema 비교, 호환성 규칙, 데이터 계약 개념까지 스키마 관리 전략을 체계적으로 학습합니다.]]></description>
      <content:encoded><![CDATA[<p>스키마 진화의 필요성, Confluent Schema Registry, Avro/Protobuf/JSON Schema 비교, 호환성 규칙, 데이터 계약 개념까지 스키마 관리 전략을 체계적으로 학습합니다.</p><p><strong>데이터</strong> / 18분 읽기 / 시리즈: realtime-pipeline</p><p><a href="https://archive.kreathlab.com/tech/realtime-pipeline-09-schema">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 26 Feb 2026 00:00:00 GMT</pubDate>
      <category>data</category>
      <category>streaming</category>
      <category>data-engineering</category>
    </item>
    <item>
      <title><![CDATA[7장: Weaviate -- 오픈소스 벡터 검색 엔진]]></title>
      <link>https://archive.kreathlab.com/tech/vector-database-07-weaviate</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/vector-database-07-weaviate</guid>
      <description><![CDATA[Weaviate의 오브젝트 지향 스키마, 모듈화 아키텍처, 내장 벡터라이저, 멀티테넌시, BlockMax WAND 하이브리드 검색, GraphQL API, 배포 옵션과 Python 실습을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Weaviate의 오브젝트 지향 스키마, 모듈화 아키텍처, 내장 벡터라이저, 멀티테넌시, BlockMax WAND 하이브리드 검색, GraphQL API, 배포 옵션과 Python 실습을 다룹니다.</p><p><strong>AI / ML</strong> / 12분 읽기 / 시리즈: vector-database</p><p><a href="https://archive.kreathlab.com/tech/vector-database-07-weaviate">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 26 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>vector-database</category>
      <category>ai</category>
      <category>embedding</category>
      <category>search</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[2장: AI-Native 애플리케이션 설계 원칙]]></title>
      <link>https://archive.kreathlab.com/tech/ai-architecture-02-ai-native-design</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-architecture-02-ai-native-design</guid>
      <description><![CDATA[AI를 핵심 컴포넌트로 설계하는 AI-Native 애플리케이션의 설계 원칙 — 분리된 추론, 폴백 전략, 비동기 처리, 그리고 인간-AI 협업 패턴을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>AI를 핵심 컴포넌트로 설계하는 AI-Native 애플리케이션의 설계 원칙 — 분리된 추론, 폴백 전략, 비동기 처리, 그리고 인간-AI 협업 패턴을 다룹니다.</p><p><strong>아키텍처</strong> / 12분 읽기 / 시리즈: ai-architecture</p><p><a href="https://archive.kreathlab.com/tech/ai-architecture-02-ai-native-design">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 25 Feb 2026 00:00:00 GMT</pubDate>
      <category>architecture</category>
      <category>architecture</category>
      <category>llm</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[6장: 커스텀 슬래시 명령어와 스킬 만들기]]></title>
      <link>https://archive.kreathlab.com/tech/claude-code-guide-06-skills</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/claude-code-guide-06-skills</guid>
      <description><![CDATA[Claude Code의 스킬 시스템을 이해하고, 반복 작업을 자동화하는 커스텀 슬래시 명령어를 설계하고 구현하는 방법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Claude Code의 스킬 시스템을 이해하고, 반복 작업을 자동화하는 커스텀 슬래시 명령어를 설계하고 구현하는 방법을 다룹니다.</p><p><strong>개발 도구</strong> / 15분 읽기 / 시리즈: claude-code-guide</p><p><a href="https://archive.kreathlab.com/tech/claude-code-guide-06-skills">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 25 Feb 2026 00:00:00 GMT</pubDate>
      <category>devtools</category>
      <category>ai-agent</category>
      <category>anthropic</category>
      <category>mcp</category>
      <category>automation</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[10장: 프로덕션 아키텍처와 최적화]]></title>
      <link>https://archive.kreathlab.com/tech/multimodal-ai-10-production</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/multimodal-ai-10-production</guid>
      <description><![CDATA[멀티모달 AI 시스템의 프로덕션 배포 전략 — 서빙 인프라, 비용 관리, 지연 시간 최적화, 캐싱, 모니터링, 그리고 확장성 설계를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>멀티모달 AI 시스템의 프로덕션 배포 전략 — 서빙 인프라, 비용 관리, 지연 시간 최적화, 캐싱, 모니터링, 그리고 확장성 설계를 다룹니다.</p><p><strong>AI / ML</strong> / 11분 읽기 / 시리즈: multimodal-ai</p><p><a href="https://archive.kreathlab.com/tech/multimodal-ai-10-production">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 25 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>multimodal</category>
      <category>embedding</category>
    </item>
    <item>
      <title><![CDATA[1장: 온디바이스 AI의 등장과 핵심 개념]]></title>
      <link>https://archive.kreathlab.com/tech/on-device-ai-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/on-device-ai-01-introduction</guid>
      <description><![CDATA[온디바이스 AI의 필요성, 클라우드 추론 대비 장단점, 핵심 기술 스택, 그리고 현재 기술 수준과 한계를 조망합니다.]]></description>
      <content:encoded><![CDATA[<p>온디바이스 AI의 필요성, 클라우드 추론 대비 장단점, 핵심 기술 스택, 그리고 현재 기술 수준과 한계를 조망합니다.</p><p><strong>AI / ML</strong> / 9분 읽기 / 시리즈: on-device-ai</p><p><a href="https://archive.kreathlab.com/tech/on-device-ai-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 25 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>performance</category>
      <category>mlops</category>
    </item>
    <item>
      <title><![CDATA[8장: 테스트와 품질 보증]]></title>
      <link>https://archive.kreathlab.com/tech/rust-backend-08-testing</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/rust-backend-08-testing</guid>
      <description><![CDATA[Rust의 단위 테스트, 통합 테스트, API 테스트, testcontainers를 활용한 DB 테스트, 프로퍼티 기반 테스트, 벤치마킹까지 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Rust의 단위 테스트, 통합 테스트, API 테스트, testcontainers를 활용한 DB 테스트, 프로퍼티 기반 테스트, 벤치마킹까지 다룹니다.</p><p><strong>프로그래밍</strong> / 12분 읽기 / 시리즈: rust-backend</p><p><a href="https://archive.kreathlab.com/tech/rust-backend-08-testing">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 25 Feb 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>rust</category>
    </item>
    <item>
      <title><![CDATA[6장: 새로운 유틸리티 클래스와 변형]]></title>
      <link>https://archive.kreathlab.com/tech/tailwind-v4-06-utilities-variants</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/tailwind-v4-06-utilities-variants</guid>
      <description><![CDATA[Tailwind CSS v4에서 추가, 변경, 제거된 유틸리티 클래스와 변형(Variants)을 종합 정리하고, 실전 활용 패턴을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Tailwind CSS v4에서 추가, 변경, 제거된 유틸리티 클래스와 변형(Variants)을 종합 정리하고, 실전 활용 패턴을 다룹니다.</p><p><strong>웹 개발</strong> / 11분 읽기 / 시리즈: tailwind-v4</p><p><a href="https://archive.kreathlab.com/tech/tailwind-v4-06-utilities-variants">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 25 Feb 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>frontend</category>
    </item>
    <item>
      <title><![CDATA[1장: AI 보안의 전체 그림과 위협 모델]]></title>
      <link>https://archive.kreathlab.com/tech/ai-security-01-threat-landscape</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-security-01-threat-landscape</guid>
      <description><![CDATA[AI 시스템이 직면하는 보안 위협의 전체 지형을 조망합니다. OWASP Top 10 for LLM, 공격 표면 분석, 위협 모델링, 그리고 방어 전략의 계층적 접근법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>AI 시스템이 직면하는 보안 위협의 전체 지형을 조망합니다. OWASP Top 10 for LLM, 공격 표면 분석, 위협 모델링, 그리고 방어 전략의 계층적 접근법을 다룹니다.</p><p><strong>AI / ML</strong> / 11분 읽기 / 시리즈: ai-security</p><p><a href="https://archive.kreathlab.com/tech/ai-security-01-threat-landscape">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 24 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>testing</category>
      <category>security</category>
    </item>
    <item>
      <title><![CDATA[11장: 실전 프로젝트 — AI 서비스 API 설계]]></title>
      <link>https://archive.kreathlab.com/tech/api-design-11-project</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/api-design-11-project</guid>
      <description><![CDATA[REST 공개 API와 gRPC 내부 통신을 결합한 AI 서비스 API를 설계하고, OpenAPI 스펙, FastAPI 구현, 스트리밍, 인증, SDK 생성까지 전체를 구축합니다.]]></description>
      <content:encoded><![CDATA[<p>REST 공개 API와 gRPC 내부 통신을 결합한 AI 서비스 API를 설계하고, OpenAPI 스펙, FastAPI 구현, 스트리밍, 인증, SDK 생성까지 전체를 구축합니다.</p><p><strong>아키텍처</strong> / 20분 읽기 / 시리즈: api-design</p><p><a href="https://archive.kreathlab.com/tech/api-design-11-project">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 24 Feb 2026 00:00:00 GMT</pubDate>
      <category>architecture</category>
      <category>api-design</category>
      <category>graphql</category>
      <category>architecture</category>
    </item>
    <item>
      <title><![CDATA[10장: 프로덕션 운영과 모니터링]]></title>
      <link>https://archive.kreathlab.com/tech/graphql-architecture-10-production</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/graphql-architecture-10-production</guid>
      <description><![CDATA[GraphQL API의 프로덕션 운영 전략을 다룹니다. 스키마 변경 관리, Breaking Change 감지, 스키마 레지스트리, 쿼리 제한, 관측 가능성, 에러 처리, 성능 모니터링을 학습합니다.]]></description>
      <content:encoded><![CDATA[<p>GraphQL API의 프로덕션 운영 전략을 다룹니다. 스키마 변경 관리, Breaking Change 감지, 스키마 레지스트리, 쿼리 제한, 관측 가능성, 에러 처리, 성능 모니터링을 학습합니다.</p><p><strong>웹 개발</strong> / 13분 읽기 / 시리즈: graphql-architecture</p><p><a href="https://archive.kreathlab.com/tech/graphql-architecture-10-production">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 24 Feb 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>graphql</category>
      <category>api-design</category>
      <category>frontend</category>
    </item>
    <item>
      <title><![CDATA[9장: AI 서비스 관측 가능성]]></title>
      <link>https://archive.kreathlab.com/tech/opentelemetry-09-ai-observability</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/opentelemetry-09-ai-observability</guid>
      <description><![CDATA[LLM 호출 추적, 토큰 사용량/비용 모니터링, AI 에이전트 행동 추적, LangChain/LlamaIndex OTel 통합을 통한 AI 관측 가능성을 학습합니다.]]></description>
      <content:encoded><![CDATA[<p>LLM 호출 추적, 토큰 사용량/비용 모니터링, AI 에이전트 행동 추적, LangChain/LlamaIndex OTel 통합을 통한 AI 관측 가능성을 학습합니다.</p><p><strong>인프라</strong> / 12분 읽기 / 시리즈: opentelemetry</p><p><a href="https://archive.kreathlab.com/tech/opentelemetry-09-ai-observability">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 24 Feb 2026 00:00:00 GMT</pubDate>
      <category>infra</category>
      <category>monitoring</category>
      <category>observability</category>
    </item>
    <item>
      <title><![CDATA[8장: CDC(Change Data Capture)]]></title>
      <link>https://archive.kreathlab.com/tech/realtime-pipeline-08-cdc</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/realtime-pipeline-08-cdc</guid>
      <description><![CDATA[CDC의 원리와 WAL 기반 변경 캡처, Debezium 아키텍처, PostgreSQL/MySQL CDC 실습, Flink CDC 3.6, 아웃박스 패턴, 이벤추얼 컨시스턴시까지 데이터 통합의 핵심을 학습합니다.]]></description>
      <content:encoded><![CDATA[<p>CDC의 원리와 WAL 기반 변경 캡처, Debezium 아키텍처, PostgreSQL/MySQL CDC 실습, Flink CDC 3.6, 아웃박스 패턴, 이벤추얼 컨시스턴시까지 데이터 통합의 핵심을 학습합니다.</p><p><strong>데이터</strong> / 18분 읽기 / 시리즈: realtime-pipeline</p><p><a href="https://archive.kreathlab.com/tech/realtime-pipeline-08-cdc">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 24 Feb 2026 00:00:00 GMT</pubDate>
      <category>data</category>
      <category>streaming</category>
      <category>data-engineering</category>
    </item>
    <item>
      <title><![CDATA[6장: Pinecone -- 매니지드 벡터 데이터베이스]]></title>
      <link>https://archive.kreathlab.com/tech/vector-database-06-pinecone</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/vector-database-06-pinecone</guid>
      <description><![CDATA[Pinecone의 완전 관리형 아키텍처, 서버리스와 팟 배포 모델, 네임스페이스, 메타데이터 필터링, 하이브리드 검색, 보안 컴플라이언스, Python SDK 실습을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Pinecone의 완전 관리형 아키텍처, 서버리스와 팟 배포 모델, 네임스페이스, 메타데이터 필터링, 하이브리드 검색, 보안 컴플라이언스, Python SDK 실습을 다룹니다.</p><p><strong>AI / ML</strong> / 13분 읽기 / 시리즈: vector-database</p><p><a href="https://archive.kreathlab.com/tech/vector-database-06-pinecone">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 24 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>vector-database</category>
      <category>ai</category>
      <category>embedding</category>
      <category>search</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[1장: AI 시대의 소프트웨어 아키텍처 — 새로운 패러다임]]></title>
      <link>https://archive.kreathlab.com/tech/ai-architecture-01-new-paradigm</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-architecture-01-new-paradigm</guid>
      <description><![CDATA[AI 통합이 소프트웨어 아키텍처에 가져온 근본적 변화, 결정론에서 확률론으로의 전환, 그리고 AI-Native 시스템의 핵심 특성을 조망합니다.]]></description>
      <content:encoded><![CDATA[<p>AI 통합이 소프트웨어 아키텍처에 가져온 근본적 변화, 결정론에서 확률론으로의 전환, 그리고 AI-Native 시스템의 핵심 특성을 조망합니다.</p><p><strong>아키텍처</strong> / 13분 읽기 / 시리즈: ai-architecture</p><p><a href="https://archive.kreathlab.com/tech/ai-architecture-01-new-paradigm">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 23 Feb 2026 00:00:00 GMT</pubDate>
      <category>architecture</category>
      <category>architecture</category>
      <category>llm</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[11장: 실전 프로젝트 — AI 검색 시스템 구축]]></title>
      <link>https://archive.kreathlab.com/tech/ai-search-11-project</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-search-11-project</guid>
      <description><![CDATA[Elasticsearch, Cross-encoder 리랭킹, 개인화를 통합한 AI 검색 시스템의 전체 아키텍처 설계부터 구현, 벤치마킹, 운영 체크리스트까지 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Elasticsearch, Cross-encoder 리랭킹, 개인화를 통합한 AI 검색 시스템의 전체 아키텍처 설계부터 구현, 벤치마킹, 운영 체크리스트까지 다룹니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: ai-search</p><p><a href="https://archive.kreathlab.com/tech/ai-search-11-project">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 23 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>search</category>
      <category>ai</category>
    </item>
    <item>
      <title><![CDATA[5장: 훅(Hooks)으로 자동화 파이프라인 구축하기]]></title>
      <link>https://archive.kreathlab.com/tech/claude-code-guide-05-hooks</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/claude-code-guide-05-hooks</guid>
      <description><![CDATA[Claude Code 훅 시스템의 개념과 라이프사이클을 이해하고, 코드 품질 게이트와 자동화 파이프라인을 구축하는 방법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Claude Code 훅 시스템의 개념과 라이프사이클을 이해하고, 코드 품질 게이트와 자동화 파이프라인을 구축하는 방법을 다룹니다.</p><p><strong>개발 도구</strong> / 13분 읽기 / 시리즈: claude-code-guide</p><p><a href="https://archive.kreathlab.com/tech/claude-code-guide-05-hooks">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 23 Feb 2026 00:00:00 GMT</pubDate>
      <category>devtools</category>
      <category>ai-agent</category>
      <category>anthropic</category>
      <category>mcp</category>
      <category>automation</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[9장: 멀티모달 에이전트 구축]]></title>
      <link>https://archive.kreathlab.com/tech/multimodal-ai-09-multimodal-agents</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/multimodal-ai-09-multimodal-agents</guid>
      <description><![CDATA[시각적 이해 능력을 갖춘 AI 에이전트의 설계와 구현 — 화면 상호작용 에이전트, 멀티모달 도구 호출, Computer Use, 그리고 실전 에이전트 패턴을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>시각적 이해 능력을 갖춘 AI 에이전트의 설계와 구현 — 화면 상호작용 에이전트, 멀티모달 도구 호출, Computer Use, 그리고 실전 에이전트 패턴을 다룹니다.</p><p><strong>AI / ML</strong> / 12분 읽기 / 시리즈: multimodal-ai</p><p><a href="https://archive.kreathlab.com/tech/multimodal-ai-09-multimodal-agents">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 23 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>multimodal</category>
      <category>embedding</category>
    </item>
    <item>
      <title><![CDATA[7장: SQLx 데이터베이스 연동]]></title>
      <link>https://archive.kreathlab.com/tech/rust-backend-07-database</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/rust-backend-07-database</guid>
      <description><![CDATA[SQLx의 컴파일 타임 쿼리 검증, 연결 풀, CRUD 구현, 마이그레이션, 트랜잭션까지 Rust 백엔드의 데이터베이스 연동 패턴을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>SQLx의 컴파일 타임 쿼리 검증, 연결 풀, CRUD 구현, 마이그레이션, 트랜잭션까지 Rust 백엔드의 데이터베이스 연동 패턴을 다룹니다.</p><p><strong>프로그래밍</strong> / 11분 읽기 / 시리즈: rust-backend</p><p><a href="https://archive.kreathlab.com/tech/rust-backend-07-database">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 23 Feb 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>rust</category>
    </item>
    <item>
      <title><![CDATA[5장: 컨테이너 쿼리와 반응형 디자인의 진화]]></title>
      <link>https://archive.kreathlab.com/tech/tailwind-v4-05-container-queries</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/tailwind-v4-05-container-queries</guid>
      <description><![CDATA[Tailwind CSS v4의 네이티브 컨테이너 쿼리 지원, @container 변형, 기존 미디어 쿼리와의 비교, 그리고 컴포넌트 기반 반응형 디자인 패턴을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Tailwind CSS v4의 네이티브 컨테이너 쿼리 지원, @container 변형, 기존 미디어 쿼리와의 비교, 그리고 컴포넌트 기반 반응형 디자인 패턴을 다룹니다.</p><p><strong>웹 개발</strong> / 11분 읽기 / 시리즈: tailwind-v4</p><p><a href="https://archive.kreathlab.com/tech/tailwind-v4-05-container-queries">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 23 Feb 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>frontend</category>
    </item>
    <item>
      <title><![CDATA[11장: 프레임워크 선택 기준과 마이그레이션 전략]]></title>
      <link>https://archive.kreathlab.com/tech/ai-orchestration-11-selection-migration</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-orchestration-11-selection-migration</guid>
      <description><![CDATA[5대 프레임워크 종합 비교, 의사결정 트리, 하이브리드 아키텍처, 마이그레이션 가이드, 프레임워크 독립적 설계 원칙을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>5대 프레임워크 종합 비교, 의사결정 트리, 하이브리드 아키텍처, 마이그레이션 가이드, 프레임워크 독립적 설계 원칙을 다룹니다.</p><p><strong>AI / ML</strong> / 20분 읽기 / 시리즈: ai-orchestration</p><p><a href="https://archive.kreathlab.com/tech/ai-orchestration-11-selection-migration">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 22 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>orchestration</category>
      <category>ai-framework</category>
      <category>ai</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[10장: API 게이트웨이와 프로덕션 인프라]]></title>
      <link>https://archive.kreathlab.com/tech/api-design-10-gateway</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/api-design-10-gateway</guid>
      <description><![CDATA[LLM 게이트웨이를 활용한 멀티 프로바이더 라우팅, 모델 폴백, 인증/인가, 캐싱, 관측 가능성 등 프로덕션 API 인프라를 학습합니다.]]></description>
      <content:encoded><![CDATA[<p>LLM 게이트웨이를 활용한 멀티 프로바이더 라우팅, 모델 폴백, 인증/인가, 캐싱, 관측 가능성 등 프로덕션 API 인프라를 학습합니다.</p><p><strong>아키텍처</strong> / 16분 읽기 / 시리즈: api-design</p><p><a href="https://archive.kreathlab.com/tech/api-design-10-gateway">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 22 Feb 2026 00:00:00 GMT</pubDate>
      <category>architecture</category>
      <category>api-design</category>
      <category>graphql</category>
      <category>architecture</category>
    </item>
    <item>
      <title><![CDATA[9장: 코드 생성과 타입 안전성]]></title>
      <link>https://archive.kreathlab.com/tech/graphql-architecture-09-codegen</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/graphql-architecture-09-codegen</guid>
      <description><![CDATA[GraphQL Code Generator를 활용한 TypeScript 타입 생성, React 훅 자동 생성, Fragment Colocation 패턴을 다룹니다. 코드 퍼스트 스키마 빌더와의 비교, CI 통합 전략도 학습합니다.]]></description>
      <content:encoded><![CDATA[<p>GraphQL Code Generator를 활용한 TypeScript 타입 생성, React 훅 자동 생성, Fragment Colocation 패턴을 다룹니다. 코드 퍼스트 스키마 빌더와의 비교, CI 통합 전략도 학습합니다.</p><p><strong>웹 개발</strong> / 13분 읽기 / 시리즈: graphql-architecture</p><p><a href="https://archive.kreathlab.com/tech/graphql-architecture-09-codegen">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 22 Feb 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>graphql</category>
      <category>api-design</category>
      <category>frontend</category>
    </item>
    <item>
      <title><![CDATA[8장: Grafana, Jaeger, Prometheus 연동]]></title>
      <link>https://archive.kreathlab.com/tech/opentelemetry-08-backends</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/opentelemetry-08-backends</guid>
      <description><![CDATA[Jaeger로 분산 추적을 시각화하고, Prometheus로 메트릭을 저장/쿼리하며, Grafana로 통합 대시보드를 구성합니다. Docker Compose로 전체 스택을 실습합니다.]]></description>
      <content:encoded><![CDATA[<p>Jaeger로 분산 추적을 시각화하고, Prometheus로 메트릭을 저장/쿼리하며, Grafana로 통합 대시보드를 구성합니다. Docker Compose로 전체 스택을 실습합니다.</p><p><strong>인프라</strong> / 12분 읽기 / 시리즈: opentelemetry</p><p><a href="https://archive.kreathlab.com/tech/opentelemetry-08-backends">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 22 Feb 2026 00:00:00 GMT</pubDate>
      <category>infra</category>
      <category>monitoring</category>
      <category>observability</category>
    </item>
    <item>
      <title><![CDATA[7장: Spark Structured Streaming]]></title>
      <link>https://archive.kreathlab.com/tech/realtime-pipeline-07-spark-streaming</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/realtime-pipeline-07-spark-streaming</guid>
      <description><![CDATA[Spark Structured Streaming의 마이크로배치와 연속 처리 모드, DataFrame API 기반 스트리밍, 윈도우와 워터마크, Kafka 소스/싱크 연동, Flink와의 비교를 학습합니다.]]></description>
      <content:encoded><![CDATA[<p>Spark Structured Streaming의 마이크로배치와 연속 처리 모드, DataFrame API 기반 스트리밍, 윈도우와 워터마크, Kafka 소스/싱크 연동, Flink와의 비교를 학습합니다.</p><p><strong>데이터</strong> / 14분 읽기 / 시리즈: realtime-pipeline</p><p><a href="https://archive.kreathlab.com/tech/realtime-pipeline-07-spark-streaming">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 22 Feb 2026 00:00:00 GMT</pubDate>
      <category>data</category>
      <category>streaming</category>
      <category>data-engineering</category>
    </item>
    <item>
      <title><![CDATA[5장: DiskANN과 대규모 인덱싱 전략]]></title>
      <link>https://archive.kreathlab.com/tech/vector-database-05-diskann</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/vector-database-05-diskann</guid>
      <description><![CDATA[DiskANN의 Vamana 그래프 아키텍처와 SSD 최적화 전략을 분석하고, 10억+ 벡터 스케일에서의 성능, Fresh DiskANN과 Filtered DiskANN을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>DiskANN의 Vamana 그래프 아키텍처와 SSD 최적화 전략을 분석하고, 10억+ 벡터 스케일에서의 성능, Fresh DiskANN과 Filtered DiskANN을 다룹니다.</p><p><strong>AI / ML</strong> / 15분 읽기 / 시리즈: vector-database</p><p><a href="https://archive.kreathlab.com/tech/vector-database-05-diskann">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 22 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>vector-database</category>
      <category>ai</category>
      <category>embedding</category>
      <category>search</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[11장: 보안 고려사항과 거버넌스]]></title>
      <link>https://archive.kreathlab.com/tech/ai-coding-assistant-11-security</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-coding-assistant-11-security</guid>
      <description><![CDATA[AI 코드 보안 취약점(40-62%), 코드 유출 위험, IP/라이선스 문제, 보안 스캐닝 통합, 거버넌스 프레임워크, 안전한 AI 코딩 워크플로우를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>AI 코드 보안 취약점(40-62%), 코드 유출 위험, IP/라이선스 문제, 보안 스캐닝 통합, 거버넌스 프레임워크, 안전한 AI 코딩 워크플로우를 다룹니다.</p><p><strong>개발 도구</strong> / 22분 읽기 / 시리즈: ai-coding-assistant</p><p><a href="https://archive.kreathlab.com/tech/ai-coding-assistant-11-security">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 21 Feb 2026 00:00:00 GMT</pubDate>
      <category>devtools</category>
      <category>ai-coding</category>
      <category>claude-code</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[10장: 피드백 루프와 지속적 개선]]></title>
      <link>https://archive.kreathlab.com/tech/ai-search-10-feedback</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-search-10-feedback</guid>
      <description><![CDATA[클릭 신호 수집, 암묵적/명시적 피드백, 온라인 학습, A/B 테스트 자동화, 검색 품질 모니터링, 차가운 시작 문제를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>클릭 신호 수집, 암묵적/명시적 피드백, 온라인 학습, A/B 테스트 자동화, 검색 품질 모니터링, 차가운 시작 문제를 다룹니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: ai-search</p><p><a href="https://archive.kreathlab.com/tech/ai-search-10-feedback">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 21 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>search</category>
      <category>ai</category>
    </item>
    <item>
      <title><![CDATA[4장: 핵심 워크플로우 - 코드 작성, 리뷰, 리팩터링]]></title>
      <link>https://archive.kreathlab.com/tech/claude-code-guide-04-core-workflows</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/claude-code-guide-04-core-workflows</guid>
      <description><![CDATA[Claude Code로 코드를 작성하고, 리뷰하고, 리팩터링하는 핵심 워크플로우를 실전 예제와 함께 익힙니다.]]></description>
      <content:encoded><![CDATA[<p>Claude Code로 코드를 작성하고, 리뷰하고, 리팩터링하는 핵심 워크플로우를 실전 예제와 함께 익힙니다.</p><p><strong>개발 도구</strong> / 14분 읽기 / 시리즈: claude-code-guide</p><p><a href="https://archive.kreathlab.com/tech/claude-code-guide-04-core-workflows">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 21 Feb 2026 00:00:00 GMT</pubDate>
      <category>devtools</category>
      <category>ai-agent</category>
      <category>anthropic</category>
      <category>mcp</category>
      <category>automation</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[8장: 멀티모달 RAG 시스템 설계]]></title>
      <link>https://archive.kreathlab.com/tech/multimodal-ai-08-multimodal-rag</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/multimodal-ai-08-multimodal-rag</guid>
      <description><![CDATA[텍스트, 이미지, 표, 차트 등 다양한 모달리티를 통합하는 멀티모달 RAG 시스템의 설계와 구현을 다룹니다. ColPali, 비전 기반 검색, 문서 파싱 전략을 배웁니다.]]></description>
      <content:encoded><![CDATA[<p>텍스트, 이미지, 표, 차트 등 다양한 모달리티를 통합하는 멀티모달 RAG 시스템의 설계와 구현을 다룹니다. ColPali, 비전 기반 검색, 문서 파싱 전략을 배웁니다.</p><p><strong>AI / ML</strong> / 13분 읽기 / 시리즈: multimodal-ai</p><p><a href="https://archive.kreathlab.com/tech/multimodal-ai-08-multimodal-rag">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 21 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>multimodal</category>
      <category>embedding</category>
    </item>
    <item>
      <title><![CDATA[6장: Axum 고급 패턴]]></title>
      <link>https://archive.kreathlab.com/tech/rust-backend-06-axum-advanced</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/rust-backend-06-axum-advanced</guid>
      <description><![CDATA[중첩 라우터, 인증/인가 미들웨어, 요청 검증, 웹소켓, SSE, API 테스트까지 프로덕션 수준의 Axum 패턴을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>중첩 라우터, 인증/인가 미들웨어, 요청 검증, 웹소켓, SSE, API 테스트까지 프로덕션 수준의 Axum 패턴을 다룹니다.</p><p><strong>프로그래밍</strong> / 10분 읽기 / 시리즈: rust-backend</p><p><a href="https://archive.kreathlab.com/tech/rust-backend-06-axum-advanced">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 21 Feb 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>rust</category>
    </item>
    <item>
      <title><![CDATA[4장: OKLCH 색상 시스템과 색상 유틸리티]]></title>
      <link>https://archive.kreathlab.com/tech/tailwind-v4-04-oklch-colors</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/tailwind-v4-04-oklch-colors</guid>
      <description><![CDATA[Tailwind CSS v4의 기본 색상 시스템인 OKLCH의 원리, 장점, 색상 팔레트 설계 방법, 그리고 투명도 수정자와 색상 조합 패턴을 실전 중심으로 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Tailwind CSS v4의 기본 색상 시스템인 OKLCH의 원리, 장점, 색상 팔레트 설계 방법, 그리고 투명도 수정자와 색상 조합 패턴을 실전 중심으로 다룹니다.</p><p><strong>웹 개발</strong> / 12분 읽기 / 시리즈: tailwind-v4</p><p><a href="https://archive.kreathlab.com/tech/tailwind-v4-04-oklch-colors">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 21 Feb 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>frontend</category>
    </item>
    <item>
      <title><![CDATA[10장: 에러 처리, 폴백, 관측 가능성]]></title>
      <link>https://archive.kreathlab.com/tech/ai-orchestration-10-reliability</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-orchestration-10-reliability</guid>
      <description><![CDATA[재시도 전략, 서킷 브레이커, OpenTelemetry 통합, 비용 추적, 프로덕션 모니터링까지 프로덕션 안정성 패턴을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>재시도 전략, 서킷 브레이커, OpenTelemetry 통합, 비용 추적, 프로덕션 모니터링까지 프로덕션 안정성 패턴을 다룹니다.</p><p><strong>AI / ML</strong> / 15분 읽기 / 시리즈: ai-orchestration</p><p><a href="https://archive.kreathlab.com/tech/ai-orchestration-10-reliability">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 20 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>orchestration</category>
      <category>ai-framework</category>
      <category>ai</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[9장: SDK 자동 생성과 개발자 경험]]></title>
      <link>https://archive.kreathlab.com/tech/api-design-09-sdk-generation</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/api-design-09-sdk-generation</guid>
      <description><![CDATA[OpenAPI 스펙에서 타입 안전 SDK를 자동 생성하고, API 문서화, 인터랙티브 플레이그라운드로 개발자 경험을 최적화하는 방법을 학습합니다.]]></description>
      <content:encoded><![CDATA[<p>OpenAPI 스펙에서 타입 안전 SDK를 자동 생성하고, API 문서화, 인터랙티브 플레이그라운드로 개발자 경험을 최적화하는 방법을 학습합니다.</p><p><strong>아키텍처</strong> / 13분 읽기 / 시리즈: api-design</p><p><a href="https://archive.kreathlab.com/tech/api-design-09-sdk-generation">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 20 Feb 2026 00:00:00 GMT</pubDate>
      <category>architecture</category>
      <category>api-design</category>
      <category>graphql</category>
      <category>architecture</category>
    </item>
    <item>
      <title><![CDATA[8장: 캐싱 전략]]></title>
      <link>https://archive.kreathlab.com/tech/graphql-architecture-08-caching</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/graphql-architecture-08-caching</guid>
      <description><![CDATA[GraphQL API의 다양한 캐싱 전략을 다룹니다. HTTP 캐싱의 한계, Persisted Queries를 통한 CDN 캐싱, @cacheControl 디렉티브, Redis 기반 리졸버 캐싱, 캐시 무효화 패턴을 학습합니다.]]></description>
      <content:encoded><![CDATA[<p>GraphQL API의 다양한 캐싱 전략을 다룹니다. HTTP 캐싱의 한계, Persisted Queries를 통한 CDN 캐싱, @cacheControl 디렉티브, Redis 기반 리졸버 캐싱, 캐시 무효화 패턴을 학습합니다.</p><p><strong>웹 개발</strong> / 13분 읽기 / 시리즈: graphql-architecture</p><p><a href="https://archive.kreathlab.com/tech/graphql-architecture-08-caching">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 20 Feb 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>graphql</category>
      <category>api-design</category>
      <category>frontend</category>
    </item>
    <item>
      <title><![CDATA[7장: Collector 아키텍처와 배포]]></title>
      <link>https://archive.kreathlab.com/tech/opentelemetry-07-collector</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/opentelemetry-07-collector</guid>
      <description><![CDATA[OTel Collector의 Receiver/Processor/Exporter 파이프라인, 핵심 프로세서 활용법, Kubernetes 환경에서의 DaemonSet/Deployment 배포를 학습합니다.]]></description>
      <content:encoded><![CDATA[<p>OTel Collector의 Receiver/Processor/Exporter 파이프라인, 핵심 프로세서 활용법, Kubernetes 환경에서의 DaemonSet/Deployment 배포를 학습합니다.</p><p><strong>인프라</strong> / 13분 읽기 / 시리즈: opentelemetry</p><p><a href="https://archive.kreathlab.com/tech/opentelemetry-07-collector">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 20 Feb 2026 00:00:00 GMT</pubDate>
      <category>infra</category>
      <category>monitoring</category>
      <category>observability</category>
    </item>
    <item>
      <title><![CDATA[6장: Apache Flink — 스트림 처리 엔진]]></title>
      <link>https://archive.kreathlab.com/tech/realtime-pipeline-06-flink</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/realtime-pipeline-06-flink</guid>
      <description><![CDATA[Flink의 JobManager/TaskManager 아키텍처, DataStream API, 윈도우와 조인, 상태 관리, Chandy-Lamport 체크포인팅, Flink SQL까지 스트림 처리 엔진의 핵심을 학습합니다.]]></description>
      <content:encoded><![CDATA[<p>Flink의 JobManager/TaskManager 아키텍처, DataStream API, 윈도우와 조인, 상태 관리, Chandy-Lamport 체크포인팅, Flink SQL까지 스트림 처리 엔진의 핵심을 학습합니다.</p><p><strong>데이터</strong> / 19분 읽기 / 시리즈: realtime-pipeline</p><p><a href="https://archive.kreathlab.com/tech/realtime-pipeline-06-flink">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 20 Feb 2026 00:00:00 GMT</pubDate>
      <category>data</category>
      <category>streaming</category>
      <category>data-engineering</category>
    </item>
    <item>
      <title><![CDATA[4장: IVF와 Product Quantization]]></title>
      <link>https://archive.kreathlab.com/tech/vector-database-04-ivf-pq</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/vector-database-04-ivf-pq</guid>
      <description><![CDATA[IVF 클러스터링 기반 검색과 Product Quantization의 원리를 분석하고, IVF+PQ 조합의 대규모 데이터셋 최적화 전략과 메모리-정확도 트레이드오프를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>IVF 클러스터링 기반 검색과 Product Quantization의 원리를 분석하고, IVF+PQ 조합의 대규모 데이터셋 최적화 전략과 메모리-정확도 트레이드오프를 다룹니다.</p><p><strong>AI / ML</strong> / 15분 읽기 / 시리즈: vector-database</p><p><a href="https://archive.kreathlab.com/tech/vector-database-04-ivf-pq">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 20 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>vector-database</category>
      <category>ai</category>
      <category>embedding</category>
      <category>search</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[10장: 팀 생산성 측정과 조직 도입]]></title>
      <link>https://archive.kreathlab.com/tech/ai-coding-assistant-10-productivity</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-coding-assistant-10-productivity</guid>
      <description><![CDATA[AI 코딩 도구의 생산성 측정의 함정, DORA 메트릭, AI 코드 비율 최적 범위, 조직 도입 전략, 온보딩, 가이드라인 설계, ROI 측정을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>AI 코딩 도구의 생산성 측정의 함정, DORA 메트릭, AI 코드 비율 최적 범위, 조직 도입 전략, 온보딩, 가이드라인 설계, ROI 측정을 다룹니다.</p><p><strong>개발 도구</strong> / 18분 읽기 / 시리즈: ai-coding-assistant</p><p><a href="https://archive.kreathlab.com/tech/ai-coding-assistant-10-productivity">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 19 Feb 2026 00:00:00 GMT</pubDate>
      <category>devtools</category>
      <category>ai-coding</category>
      <category>claude-code</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[9장: 검색 개인화]]></title>
      <link>https://archive.kreathlab.com/tech/ai-search-09-personalization</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-search-09-personalization</guid>
      <description><![CDATA[사용자 프로파일링, 클릭 이력 기반 개인화, 임베딩 기반 사용자 벡터, 인기도 편향 문제, 프라이버시 고려사항과 실시간 개인화를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>사용자 프로파일링, 클릭 이력 기반 개인화, 임베딩 기반 사용자 벡터, 인기도 편향 문제, 프라이버시 고려사항과 실시간 개인화를 다룹니다.</p><p><strong>AI / ML</strong> / 15분 읽기 / 시리즈: ai-search</p><p><a href="https://archive.kreathlab.com/tech/ai-search-09-personalization">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 19 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>search</category>
      <category>ai</category>
    </item>
    <item>
      <title><![CDATA[3장: CLAUDE.md로 프로젝트 컨텍스트 설계하기]]></title>
      <link>https://archive.kreathlab.com/tech/claude-code-guide-03-claude-md</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/claude-code-guide-03-claude-md</guid>
      <description><![CDATA[CLAUDE.md 파일을 활용하여 프로젝트의 기술 스택, 코딩 컨벤션, 아키텍처 정보를 Claude Code에 효과적으로 전달하는 방법을 알아봅니다.]]></description>
      <content:encoded><![CDATA[<p>CLAUDE.md 파일을 활용하여 프로젝트의 기술 스택, 코딩 컨벤션, 아키텍처 정보를 Claude Code에 효과적으로 전달하는 방법을 알아봅니다.</p><p><strong>개발 도구</strong> / 14분 읽기 / 시리즈: claude-code-guide</p><p><a href="https://archive.kreathlab.com/tech/claude-code-guide-03-claude-md">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 19 Feb 2026 00:00:00 GMT</pubDate>
      <category>devtools</category>
      <category>ai-agent</category>
      <category>anthropic</category>
      <category>mcp</category>
      <category>automation</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[7장: 멀티모달 임베딩과 크로스모달 검색]]></title>
      <link>https://archive.kreathlab.com/tech/multimodal-ai-07-embeddings-search</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/multimodal-ai-07-embeddings-search</guid>
      <description><![CDATA[CLIP 기반 멀티모달 임베딩의 원리, 텍스트-이미지 크로스모달 검색, 통합 벡터 스토어 설계, 그리고 실전 멀티모달 검색 시스템 구축을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>CLIP 기반 멀티모달 임베딩의 원리, 텍스트-이미지 크로스모달 검색, 통합 벡터 스토어 설계, 그리고 실전 멀티모달 검색 시스템 구축을 다룹니다.</p><p><strong>AI / ML</strong> / 11분 읽기 / 시리즈: multimodal-ai</p><p><a href="https://archive.kreathlab.com/tech/multimodal-ai-07-embeddings-search">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 19 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>multimodal</category>
      <category>embedding</category>
    </item>
    <item>
      <title><![CDATA[5장: Axum 웹 프레임워크 기초]]></title>
      <link>https://archive.kreathlab.com/tech/rust-backend-05-axum-basics</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/rust-backend-05-axum-basics</guid>
      <description><![CDATA[Axum의 Tower 기반 아키텍처, 라우팅, 핸들러, 추출자, 응답 타입, 미들웨어까지 Hello World에서 CRUD API까지 실습합니다.]]></description>
      <content:encoded><![CDATA[<p>Axum의 Tower 기반 아키텍처, 라우팅, 핸들러, 추출자, 응답 타입, 미들웨어까지 Hello World에서 CRUD API까지 실습합니다.</p><p><strong>프로그래밍</strong> / 10분 읽기 / 시리즈: rust-backend</p><p><a href="https://archive.kreathlab.com/tech/rust-backend-05-axum-basics">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 19 Feb 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>rust</category>
    </item>
    <item>
      <title><![CDATA[3장: 새로운 임포트 구조와 레이어 시스템]]></title>
      <link>https://archive.kreathlab.com/tech/tailwind-v4-03-layers-imports</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/tailwind-v4-03-layers-imports</guid>
      <description><![CDATA[Tailwind CSS v4의 임포트 방식, CSS 레이어 시스템, @utility로 커스텀 유틸리티 정의하기, 그리고 스타일 우선순위 관리 전략을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Tailwind CSS v4의 임포트 방식, CSS 레이어 시스템, @utility로 커스텀 유틸리티 정의하기, 그리고 스타일 우선순위 관리 전략을 다룹니다.</p><p><strong>웹 개발</strong> / 11분 읽기 / 시리즈: tailwind-v4</p><p><a href="https://archive.kreathlab.com/tech/tailwind-v4-03-layers-imports">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 19 Feb 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>frontend</category>
    </item>
    <item>
      <title><![CDATA[9장: 스트리밍과 실시간 처리 패턴]]></title>
      <link>https://archive.kreathlab.com/tech/ai-orchestration-09-streaming</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-orchestration-09-streaming</guid>
      <description><![CDATA[SSE/WebSocket, 토큰/이벤트 스트리밍, 구조화된 출력 스트리밍을 각 프레임워크별로 비교하고 프론트엔드 통합을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>SSE/WebSocket, 토큰/이벤트 스트리밍, 구조화된 출력 스트리밍을 각 프레임워크별로 비교하고 프론트엔드 통합을 다룹니다.</p><p><strong>AI / ML</strong> / 13분 읽기 / 시리즈: ai-orchestration</p><p><a href="https://archive.kreathlab.com/tech/ai-orchestration-09-streaming">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 18 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>orchestration</category>
      <category>ai-framework</category>
      <category>ai</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[8장: 레이트 리미팅과 비용 제어]]></title>
      <link>https://archive.kreathlab.com/tech/api-design-08-rate-limiting</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/api-design-08-rate-limiting</guid>
      <description><![CDATA[토큰 기반 레이트 리미팅, 토큰 버킷과 슬라이딩 윈도우 알고리즘, 사용자별 한도 설정, 비용 캡, Redis 기반 구현을 학습합니다.]]></description>
      <content:encoded><![CDATA[<p>토큰 기반 레이트 리미팅, 토큰 버킷과 슬라이딩 윈도우 알고리즘, 사용자별 한도 설정, 비용 캡, Redis 기반 구현을 학습합니다.</p><p><strong>아키텍처</strong> / 16분 읽기 / 시리즈: api-design</p><p><a href="https://archive.kreathlab.com/tech/api-design-08-rate-limiting">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 18 Feb 2026 00:00:00 GMT</pubDate>
      <category>architecture</category>
      <category>api-design</category>
      <category>graphql</category>
      <category>architecture</category>
    </item>
    <item>
      <title><![CDATA[7장: 인증과 인가]]></title>
      <link>https://archive.kreathlab.com/tech/graphql-architecture-07-auth</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/graphql-architecture-07-auth</guid>
      <description><![CDATA[GraphQL API에서의 인증과 인가 전략을 다룹니다. JWT 인증 미들웨어, 컨텍스트 기반 사용자 주입, 리졸버 수준 인가, 디렉티브 기반 선언적 인가, RBAC/ABAC 모델을 학습합니다.]]></description>
      <content:encoded><![CDATA[<p>GraphQL API에서의 인증과 인가 전략을 다룹니다. JWT 인증 미들웨어, 컨텍스트 기반 사용자 주입, 리졸버 수준 인가, 디렉티브 기반 선언적 인가, RBAC/ABAC 모델을 학습합니다.</p><p><strong>웹 개발</strong> / 13분 읽기 / 시리즈: graphql-architecture</p><p><a href="https://archive.kreathlab.com/tech/graphql-architecture-07-auth">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 18 Feb 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>graphql</category>
      <category>api-design</category>
      <category>frontend</category>
    </item>
    <item>
      <title><![CDATA[6장: OTel SDK 계측 실전]]></title>
      <link>https://archive.kreathlab.com/tech/opentelemetry-06-sdk</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/opentelemetry-06-sdk</guid>
      <description><![CDATA[자동 계측과 수동 계측의 차이를 이해하고, Python/Node.js/Go 각 언어별 SDK 활용법과 커스텀 스팬/메트릭 생성을 실습합니다.]]></description>
      <content:encoded><![CDATA[<p>자동 계측과 수동 계측의 차이를 이해하고, Python/Node.js/Go 각 언어별 SDK 활용법과 커스텀 스팬/메트릭 생성을 실습합니다.</p><p><strong>인프라</strong> / 11분 읽기 / 시리즈: opentelemetry</p><p><a href="https://archive.kreathlab.com/tech/opentelemetry-06-sdk">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 18 Feb 2026 00:00:00 GMT</pubDate>
      <category>infra</category>
      <category>monitoring</category>
      <category>observability</category>
    </item>
    <item>
      <title><![CDATA[5장: Kafka Connect와 데이터 통합]]></title>
      <link>https://archive.kreathlab.com/tech/realtime-pipeline-05-kafka-connect</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/realtime-pipeline-05-kafka-connect</guid>
      <description><![CDATA[Kafka Connect의 Source/Sink 아키텍처, 주요 커넥터 활용, 분산 모드 운영, SMT를 통한 데이터 변환, 커스텀 커넥터 개발까지 Kafka 기반 데이터 통합 전략을 학습합니다.]]></description>
      <content:encoded><![CDATA[<p>Kafka Connect의 Source/Sink 아키텍처, 주요 커넥터 활용, 분산 모드 운영, SMT를 통한 데이터 변환, 커스텀 커넥터 개발까지 Kafka 기반 데이터 통합 전략을 학습합니다.</p><p><strong>데이터</strong> / 14분 읽기 / 시리즈: realtime-pipeline</p><p><a href="https://archive.kreathlab.com/tech/realtime-pipeline-05-kafka-connect">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 18 Feb 2026 00:00:00 GMT</pubDate>
      <category>data</category>
      <category>streaming</category>
      <category>data-engineering</category>
    </item>
    <item>
      <title><![CDATA[3장: HNSW 알고리즘 심층 분석]]></title>
      <link>https://archive.kreathlab.com/tech/vector-database-03-hnsw</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/vector-database-03-hnsw</guid>
      <description><![CDATA[HNSW 알고리즘의 원리를 NSW 그래프에서부터 다층 구조까지 단계별로 분석하고, 핵심 파라미터 튜닝과 성능 특성, 적합한 사용 시나리오를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>HNSW 알고리즘의 원리를 NSW 그래프에서부터 다층 구조까지 단계별로 분석하고, 핵심 파라미터 튜닝과 성능 특성, 적합한 사용 시나리오를 다룹니다.</p><p><strong>AI / ML</strong> / 15분 읽기 / 시리즈: vector-database</p><p><a href="https://archive.kreathlab.com/tech/vector-database-03-hnsw">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 18 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>vector-database</category>
      <category>ai</category>
      <category>embedding</category>
      <category>search</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[9장: 코드 생성 품질 평가와 벤치마킹]]></title>
      <link>https://archive.kreathlab.com/tech/ai-coding-assistant-09-quality</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-coding-assistant-09-quality</guid>
      <description><![CDATA[HumanEval, SWE-bench, CursorBench 등 주요 벤치마크, pass@k 메트릭, AI 코드 품질 문제, 품질 게이트 설계, 자동화된 검증 파이프라인을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>HumanEval, SWE-bench, CursorBench 등 주요 벤치마크, pass@k 메트릭, AI 코드 품질 문제, 품질 게이트 설계, 자동화된 검증 파이프라인을 다룹니다.</p><p><strong>개발 도구</strong> / 17분 읽기 / 시리즈: ai-coding-assistant</p><p><a href="https://archive.kreathlab.com/tech/ai-coding-assistant-09-quality">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 17 Feb 2026 00:00:00 GMT</pubDate>
      <category>devtools</category>
      <category>ai-coding</category>
      <category>claude-code</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[8장: 하이브리드 검색과 리랭킹 파이프라인]]></title>
      <link>https://archive.kreathlab.com/tech/ai-search-08-hybrid-reranking</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-search-08-hybrid-reranking</guid>
      <description><![CDATA[BM25와 시맨틱 검색의 결합 전략, RRF/선형 보간, 리랭킹 캐스케이드, 다단계 검색 파이프라인 설계와 성능-품질 트레이드오프를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>BM25와 시맨틱 검색의 결합 전략, RRF/선형 보간, 리랭킹 캐스케이드, 다단계 검색 파이프라인 설계와 성능-품질 트레이드오프를 다룹니다.</p><p><strong>AI / ML</strong> / 13분 읽기 / 시리즈: ai-search</p><p><a href="https://archive.kreathlab.com/tech/ai-search-08-hybrid-reranking">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 17 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>search</category>
      <category>ai</category>
    </item>
    <item>
      <title><![CDATA[2장: 설치, 설정, 첫 실행]]></title>
      <link>https://archive.kreathlab.com/tech/claude-code-guide-02-setup-and-config</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/claude-code-guide-02-setup-and-config</guid>
      <description><![CDATA[Claude Code를 설치하고 기본 설정을 완료한 뒤, 첫 번째 대화형 세션을 실행하는 과정을 단계별로 안내합니다.]]></description>
      <content:encoded><![CDATA[<p>Claude Code를 설치하고 기본 설정을 완료한 뒤, 첫 번째 대화형 세션을 실행하는 과정을 단계별로 안내합니다.</p><p><strong>개발 도구</strong> / 12분 읽기 / 시리즈: claude-code-guide</p><p><a href="https://archive.kreathlab.com/tech/claude-code-guide-02-setup-and-config">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 17 Feb 2026 00:00:00 GMT</pubDate>
      <category>devtools</category>
      <category>ai-agent</category>
      <category>anthropic</category>
      <category>mcp</category>
      <category>automation</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[6장: 비디오 이해와 분석]]></title>
      <link>https://archive.kreathlab.com/tech/multimodal-ai-06-video-understanding</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/multimodal-ai-06-video-understanding</guid>
      <description><![CDATA[멀티모달 AI를 활용한 비디오 이해 기법 — 프레임 추출 전략, 시간적 추론, 영상 요약, 그리고 실시간 비디오 분석 파이프라인 설계를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>멀티모달 AI를 활용한 비디오 이해 기법 — 프레임 추출 전략, 시간적 추론, 영상 요약, 그리고 실시간 비디오 분석 파이프라인 설계를 다룹니다.</p><p><strong>AI / ML</strong> / 12분 읽기 / 시리즈: multimodal-ai</p><p><a href="https://archive.kreathlab.com/tech/multimodal-ai-06-video-understanding">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 17 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>multimodal</category>
      <category>embedding</category>
    </item>
    <item>
      <title><![CDATA[4장: async/await와 Tokio 런타임]]></title>
      <link>https://archive.kreathlab.com/tech/rust-backend-04-async</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/rust-backend-04-async</guid>
      <description><![CDATA[Rust의 비동기 프로그래밍 기초부터 Future 트레이트, Tokio 런타임, spawn/select/join, 채널 기반 동시성 패턴까지 체계적으로 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Rust의 비동기 프로그래밍 기초부터 Future 트레이트, Tokio 런타임, spawn/select/join, 채널 기반 동시성 패턴까지 체계적으로 다룹니다.</p><p><strong>프로그래밍</strong> / 15분 읽기 / 시리즈: rust-backend</p><p><a href="https://archive.kreathlab.com/tech/rust-backend-04-async">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 17 Feb 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>rust</category>
    </item>
    <item>
      <title><![CDATA[2장: CSS-first 설정과 @theme 디렉티브]]></title>
      <link>https://archive.kreathlab.com/tech/tailwind-v4-02-theme-directive</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/tailwind-v4-02-theme-directive</guid>
      <description><![CDATA[Tailwind CSS v4의 핵심 설정 방식인 @theme 디렉티브를 심층 분석합니다. 디자인 토큰 정의, 네임스페이스 규칙, 기본 테마 확장과 재정의 패턴을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Tailwind CSS v4의 핵심 설정 방식인 @theme 디렉티브를 심층 분석합니다. 디자인 토큰 정의, 네임스페이스 규칙, 기본 테마 확장과 재정의 패턴을 다룹니다.</p><p><strong>웹 개발</strong> / 11분 읽기 / 시리즈: tailwind-v4</p><p><a href="https://archive.kreathlab.com/tech/tailwind-v4-02-theme-directive">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 17 Feb 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>frontend</category>
    </item>
    <item>
      <title><![CDATA[8장: 메모리 관리와 상태 유지]]></title>
      <link>https://archive.kreathlab.com/tech/ai-orchestration-08-memory</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-orchestration-08-memory</guid>
      <description><![CDATA[대화 메모리, 장기 메모리, 벡터 메모리, 구조화된 상태를 각 프레임워크별로 비교하고 프로덕션 메모리 전략을 정리합니다.]]></description>
      <content:encoded><![CDATA[<p>대화 메모리, 장기 메모리, 벡터 메모리, 구조화된 상태를 각 프레임워크별로 비교하고 프로덕션 메모리 전략을 정리합니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: ai-orchestration</p><p><a href="https://archive.kreathlab.com/tech/ai-orchestration-08-memory">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 16 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>orchestration</category>
      <category>ai-framework</category>
      <category>ai</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[7장: API 버전 관리와 하위 호환성]]></title>
      <link>https://archive.kreathlab.com/tech/api-design-07-versioning</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/api-design-07-versioning</guid>
      <description><![CDATA[URL 경로, 헤더, 쿼리 파라미터 버전 관리 전략과 AI 서비스에서의 모델 버전 분리, 프롬프트 버전 관리, 폐기 정책을 학습합니다.]]></description>
      <content:encoded><![CDATA[<p>URL 경로, 헤더, 쿼리 파라미터 버전 관리 전략과 AI 서비스에서의 모델 버전 분리, 프롬프트 버전 관리, 폐기 정책을 학습합니다.</p><p><strong>아키텍처</strong> / 16분 읽기 / 시리즈: api-design</p><p><a href="https://archive.kreathlab.com/tech/api-design-07-versioning">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 16 Feb 2026 00:00:00 GMT</pubDate>
      <category>architecture</category>
      <category>api-design</category>
      <category>graphql</category>
      <category>architecture</category>
    </item>
    <item>
      <title><![CDATA[6장: 실시간 구독(Subscriptions)]]></title>
      <link>https://archive.kreathlab.com/tech/graphql-architecture-06-subscriptions</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/graphql-architecture-06-subscriptions</guid>
      <description><![CDATA[GraphQL Subscriptions의 프로토콜과 구현을 다룹니다. WebSocket 전송, PubSub 패턴, Redis를 활용한 확장, 필터링 전략, 그리고 Federation 환경에서의 한계와 SSE 등 대안 패턴을 살펴봅니다.]]></description>
      <content:encoded><![CDATA[<p>GraphQL Subscriptions의 프로토콜과 구현을 다룹니다. WebSocket 전송, PubSub 패턴, Redis를 활용한 확장, 필터링 전략, 그리고 Federation 환경에서의 한계와 SSE 등 대안 패턴을 살펴봅니다.</p><p><strong>웹 개발</strong> / 13분 읽기 / 시리즈: graphql-architecture</p><p><a href="https://archive.kreathlab.com/tech/graphql-architecture-06-subscriptions">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 16 Feb 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>graphql</category>
      <category>api-design</category>
      <category>frontend</category>
    </item>
    <item>
      <title><![CDATA[5장: 로그 통합과 상관관계]]></title>
      <link>https://archive.kreathlab.com/tech/opentelemetry-05-logs</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/opentelemetry-05-logs</guid>
      <description><![CDATA[OpenTelemetry 로그 데이터 모델, 로그-트레이스 상관관계, 기존 로거 브릿지(Python logging, Go slog), 구조화 로그와 로그 레벨 전략을 학습합니다.]]></description>
      <content:encoded><![CDATA[<p>OpenTelemetry 로그 데이터 모델, 로그-트레이스 상관관계, 기존 로거 브릿지(Python logging, Go slog), 구조화 로그와 로그 레벨 전략을 학습합니다.</p><p><strong>인프라</strong> / 14분 읽기 / 시리즈: opentelemetry</p><p><a href="https://archive.kreathlab.com/tech/opentelemetry-05-logs">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 16 Feb 2026 00:00:00 GMT</pubDate>
      <category>infra</category>
      <category>monitoring</category>
      <category>observability</category>
    </item>
    <item>
      <title><![CDATA[4장: Kafka 프로듀서와 컨슈머 고급 패턴]]></title>
      <link>https://archive.kreathlab.com/tech/realtime-pipeline-04-kafka-advanced</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/realtime-pipeline-04-kafka-advanced</guid>
      <description><![CDATA[Idempotent 프로듀서, 트랜잭셔널 프로듀서, Exactly-once 시맨틱스, 수동 오프셋 관리, 배치 최적화, Dead Letter Queue 등 프로덕션 수준의 Kafka 활용 패턴을 학습합니다.]]></description>
      <content:encoded><![CDATA[<p>Idempotent 프로듀서, 트랜잭셔널 프로듀서, Exactly-once 시맨틱스, 수동 오프셋 관리, 배치 최적화, Dead Letter Queue 등 프로덕션 수준의 Kafka 활용 패턴을 학습합니다.</p><p><strong>데이터</strong> / 18분 읽기 / 시리즈: realtime-pipeline</p><p><a href="https://archive.kreathlab.com/tech/realtime-pipeline-04-kafka-advanced">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 16 Feb 2026 00:00:00 GMT</pubDate>
      <category>data</category>
      <category>streaming</category>
      <category>data-engineering</category>
    </item>
    <item>
      <title><![CDATA[2장: 벡터 임베딩과 유사도 검색 기초]]></title>
      <link>https://archive.kreathlab.com/tech/vector-database-02-embeddings</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/vector-database-02-embeddings</guid>
      <description><![CDATA[임베딩의 원리와 텍스트, 이미지, 멀티모달 임베딩 모델을 비교하고, 유사도 메트릭의 수학적 배경과 차원의 저주, 임베딩 모델 선택 가이드를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>임베딩의 원리와 텍스트, 이미지, 멀티모달 임베딩 모델을 비교하고, 유사도 메트릭의 수학적 배경과 차원의 저주, 임베딩 모델 선택 가이드를 다룹니다.</p><p><strong>AI / ML</strong> / 15분 읽기 / 시리즈: vector-database</p><p><a href="https://archive.kreathlab.com/tech/vector-database-02-embeddings">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 16 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>vector-database</category>
      <category>ai</category>
      <category>embedding</category>
      <category>search</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[8장: 프롬프트 최적화와 효과적 사용법]]></title>
      <link>https://archive.kreathlab.com/tech/ai-coding-assistant-08-prompting</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-coding-assistant-08-prompting</guid>
      <description><![CDATA[코딩 프롬프트 패턴, 작업 분해 전략, 반복 개선 워크플로우, 코드 리뷰/디버깅/리팩터링 프롬프팅, 도구별 최적 사용법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>코딩 프롬프트 패턴, 작업 분해 전략, 반복 개선 워크플로우, 코드 리뷰/디버깅/리팩터링 프롬프팅, 도구별 최적 사용법을 다룹니다.</p><p><strong>개발 도구</strong> / 18분 읽기 / 시리즈: ai-coding-assistant</p><p><a href="https://archive.kreathlab.com/tech/ai-coding-assistant-08-prompting">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 15 Feb 2026 00:00:00 GMT</pubDate>
      <category>devtools</category>
      <category>ai-coding</category>
      <category>claude-code</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[7장: OpenSearch와 기타 검색 엔진]]></title>
      <link>https://archive.kreathlab.com/tech/ai-search-07-opensearch</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-search-07-opensearch</guid>
      <description><![CDATA[OpenSearch 신경 검색, 재랭킹 파이프라인과 Algolia, Meilisearch, Typesense 등 주요 검색 엔진의 AI 검색 기능을 비교합니다.]]></description>
      <content:encoded><![CDATA[<p>OpenSearch 신경 검색, 재랭킹 파이프라인과 Algolia, Meilisearch, Typesense 등 주요 검색 엔진의 AI 검색 기능을 비교합니다.</p><p><strong>AI / ML</strong> / 12분 읽기 / 시리즈: ai-search</p><p><a href="https://archive.kreathlab.com/tech/ai-search-07-opensearch">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 15 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>search</category>
      <category>ai</category>
    </item>
    <item>
      <title><![CDATA[1장: Claude Code 소개와 핵심 개념]]></title>
      <link>https://archive.kreathlab.com/tech/claude-code-guide-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/claude-code-guide-01-introduction</guid>
      <description><![CDATA[Claude Code가 무엇이고 왜 필요한지, 기존 AI 코딩 도구와 어떻게 다른지 핵심 아키텍처와 함께 살펴봅니다.]]></description>
      <content:encoded><![CDATA[<p>Claude Code가 무엇이고 왜 필요한지, 기존 AI 코딩 도구와 어떻게 다른지 핵심 아키텍처와 함께 살펴봅니다.</p><p><strong>개발 도구</strong> / 14분 읽기 / 시리즈: claude-code-guide</p><p><a href="https://archive.kreathlab.com/tech/claude-code-guide-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 15 Feb 2026 00:00:00 GMT</pubDate>
      <category>devtools</category>
      <category>ai-agent</category>
      <category>anthropic</category>
      <category>mcp</category>
      <category>automation</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[5장: 음성 AI — STT, TTS, 실시간 음성 대화]]></title>
      <link>https://archive.kreathlab.com/tech/multimodal-ai-05-speech-ai</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/multimodal-ai-05-speech-ai</guid>
      <description><![CDATA[음성 인식(STT), 음성 합성(TTS), 실시간 음성 대화 시스템의 원리와 구현을 다룹니다. Whisper, OpenAI Audio API, 음성 에이전트 설계 패턴을 배웁니다.]]></description>
      <content:encoded><![CDATA[<p>음성 인식(STT), 음성 합성(TTS), 실시간 음성 대화 시스템의 원리와 구현을 다룹니다. Whisper, OpenAI Audio API, 음성 에이전트 설계 패턴을 배웁니다.</p><p><strong>AI / ML</strong> / 11분 읽기 / 시리즈: multimodal-ai</p><p><a href="https://archive.kreathlab.com/tech/multimodal-ai-05-speech-ai">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 15 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>multimodal</category>
      <category>embedding</category>
    </item>
    <item>
      <title><![CDATA[3장: Rust 타입 시스템과 에러 처리]]></title>
      <link>https://archive.kreathlab.com/tech/rust-backend-03-type-system</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/rust-backend-03-type-system</guid>
      <description><![CDATA[구조체, 열거형, 트레이트부터 Result/Option, thiserror/anyhow를 활용한 에러 처리 패턴까지 Rust 타입 시스템의 핵심을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>구조체, 열거형, 트레이트부터 Result/Option, thiserror/anyhow를 활용한 에러 처리 패턴까지 Rust 타입 시스템의 핵심을 다룹니다.</p><p><strong>프로그래밍</strong> / 14분 읽기 / 시리즈: rust-backend</p><p><a href="https://archive.kreathlab.com/tech/rust-backend-03-type-system">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 15 Feb 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>rust</category>
    </item>
    <item>
      <title><![CDATA[1장: Tailwind CSS v4 소개 — 패러다임의 전환]]></title>
      <link>https://archive.kreathlab.com/tech/tailwind-v4-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/tailwind-v4-01-introduction</guid>
      <description><![CDATA[Tailwind CSS v4의 핵심 변경사항과 아키텍처 혁신을 조망합니다. Rust 기반 엔진, CSS-first 설정, 새로운 임포트 구조 등 v3에서 v4로의 패러다임 전환을 이해합니다.]]></description>
      <content:encoded><![CDATA[<p>Tailwind CSS v4의 핵심 변경사항과 아키텍처 혁신을 조망합니다. Rust 기반 엔진, CSS-first 설정, 새로운 임포트 구조 등 v3에서 v4로의 패러다임 전환을 이해합니다.</p><p><strong>웹 개발</strong> / 11분 읽기 / 시리즈: tailwind-v4</p><p><a href="https://archive.kreathlab.com/tech/tailwind-v4-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 15 Feb 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>frontend</category>
    </item>
    <item>
      <title><![CDATA[7장: 체이닝과 라우팅 패턴]]></title>
      <link>https://archive.kreathlab.com/tech/ai-orchestration-07-chaining-routing</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-orchestration-07-chaining-routing</guid>
      <description><![CDATA[순차/병렬 체이닝, 조건부/시맨틱 라우팅, 폴백 체인을 각 프레임워크별로 비교 구현하며 실전 패턴을 정리합니다.]]></description>
      <content:encoded><![CDATA[<p>순차/병렬 체이닝, 조건부/시맨틱 라우팅, 폴백 체인을 각 프레임워크별로 비교 구현하며 실전 패턴을 정리합니다.</p><p><strong>AI / ML</strong> / 15분 읽기 / 시리즈: ai-orchestration</p><p><a href="https://archive.kreathlab.com/tech/ai-orchestration-07-chaining-routing">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 14 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>orchestration</category>
      <category>ai-framework</category>
      <category>ai</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[6장: 스트리밍 응답 인터페이스 설계]]></title>
      <link>https://archive.kreathlab.com/tech/api-design-06-streaming</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/api-design-06-streaming</guid>
      <description><![CDATA[SSE 기반 토큰 스트리밍 프로토콜, OpenAI 호환 스트리밍 형식, 에러 처리, 클라이언트 취소, 프론트엔드 통합 패턴을 학습합니다.]]></description>
      <content:encoded><![CDATA[<p>SSE 기반 토큰 스트리밍 프로토콜, OpenAI 호환 스트리밍 형식, 에러 처리, 클라이언트 취소, 프론트엔드 통합 패턴을 학습합니다.</p><p><strong>아키텍처</strong> / 15분 읽기 / 시리즈: api-design</p><p><a href="https://archive.kreathlab.com/tech/api-design-06-streaming">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 14 Feb 2026 00:00:00 GMT</pubDate>
      <category>architecture</category>
      <category>api-design</category>
      <category>graphql</category>
      <category>architecture</category>
    </item>
    <item>
      <title><![CDATA[5장: Federation과 Supergraph]]></title>
      <link>https://archive.kreathlab.com/tech/graphql-architecture-05-federation</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/graphql-architecture-05-federation</guid>
      <description><![CDATA[Apollo Federation 2.0을 활용한 마이크로서비스 GraphQL 아키텍처를 다룹니다. 서브그래프와 슈퍼그래프 개념, 엔티티 참조 리졸버, 주요 디렉티브, 스키마 컴포지션 과정을 학습합니다.]]></description>
      <content:encoded><![CDATA[<p>Apollo Federation 2.0을 활용한 마이크로서비스 GraphQL 아키텍처를 다룹니다. 서브그래프와 슈퍼그래프 개념, 엔티티 참조 리졸버, 주요 디렉티브, 스키마 컴포지션 과정을 학습합니다.</p><p><strong>웹 개발</strong> / 15분 읽기 / 시리즈: graphql-architecture</p><p><a href="https://archive.kreathlab.com/tech/graphql-architecture-05-federation">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 14 Feb 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>graphql</category>
      <category>api-design</category>
      <category>frontend</category>
    </item>
    <item>
      <title><![CDATA[4장: 메트릭(Metrics) 수집과 분석]]></title>
      <link>https://archive.kreathlab.com/tech/opentelemetry-04-metrics</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/opentelemetry-04-metrics</guid>
      <description><![CDATA[OpenTelemetry 메트릭의 종류(Counter, Gauge, Histogram), 카디널리티 관리, Exemplars를 통한 메트릭-트레이스 연결, Prometheus 호환을 학습합니다.]]></description>
      <content:encoded><![CDATA[<p>OpenTelemetry 메트릭의 종류(Counter, Gauge, Histogram), 카디널리티 관리, Exemplars를 통한 메트릭-트레이스 연결, Prometheus 호환을 학습합니다.</p><p><strong>인프라</strong> / 14분 읽기 / 시리즈: opentelemetry</p><p><a href="https://archive.kreathlab.com/tech/opentelemetry-04-metrics">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 14 Feb 2026 00:00:00 GMT</pubDate>
      <category>infra</category>
      <category>monitoring</category>
      <category>observability</category>
    </item>
    <item>
      <title><![CDATA[3장: Apache Kafka 심층 분석]]></title>
      <link>https://archive.kreathlab.com/tech/realtime-pipeline-03-kafka-core</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/realtime-pipeline-03-kafka-core</guid>
      <description><![CDATA[Kafka의 핵심 아키텍처를 심층적으로 분석합니다. KRaft 모드, 브로커와 파티션 레플리케이션, 프로듀서 전송 보장, 컨슈머 그룹과 리밸런싱까지 Kafka의 내부를 이해합니다.]]></description>
      <content:encoded><![CDATA[<p>Kafka의 핵심 아키텍처를 심층적으로 분석합니다. KRaft 모드, 브로커와 파티션 레플리케이션, 프로듀서 전송 보장, 컨슈머 그룹과 리밸런싱까지 Kafka의 내부를 이해합니다.</p><p><strong>데이터</strong> / 18분 읽기 / 시리즈: realtime-pipeline</p><p><a href="https://archive.kreathlab.com/tech/realtime-pipeline-03-kafka-core">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 14 Feb 2026 00:00:00 GMT</pubDate>
      <category>data</category>
      <category>streaming</category>
      <category>data-engineering</category>
    </item>
    <item>
      <title><![CDATA[1장: 벡터 데이터베이스의 등장과 핵심 개념]]></title>
      <link>https://archive.kreathlab.com/tech/vector-database-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/vector-database-01-introduction</guid>
      <description><![CDATA[벡터 데이터베이스가 등장한 배경과 전통 데이터베이스와의 차이점, 유사도 검색 원리, ANN 알고리즘의 필요성, 그리고 주요 벡터 데이터베이스 생태계를 개괄합니다.]]></description>
      <content:encoded><![CDATA[<p>벡터 데이터베이스가 등장한 배경과 전통 데이터베이스와의 차이점, 유사도 검색 원리, ANN 알고리즘의 필요성, 그리고 주요 벡터 데이터베이스 생태계를 개괄합니다.</p><p><strong>AI / ML</strong> / 14분 읽기 / 시리즈: vector-database</p><p><a href="https://archive.kreathlab.com/tech/vector-database-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 14 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>vector-database</category>
      <category>ai</category>
      <category>embedding</category>
      <category>search</category>
      <category>infrastructure</category>
    </item>
    <item>
      <title><![CDATA[7장: 컨텍스트 윈도우 관리와 최적화]]></title>
      <link>https://archive.kreathlab.com/tech/ai-coding-assistant-07-context</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-coding-assistant-07-context</guid>
      <description><![CDATA[컨텍스트 엔지니어링의 부상, 파일 선택 전략, 프로젝트 규칙 파일, RAG 기반 코드 검색, 1M 토큰 시대의 전략을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>컨텍스트 엔지니어링의 부상, 파일 선택 전략, 프로젝트 규칙 파일, RAG 기반 코드 검색, 1M 토큰 시대의 전략을 다룹니다.</p><p><strong>개발 도구</strong> / 19분 읽기 / 시리즈: ai-coding-assistant</p><p><a href="https://archive.kreathlab.com/tech/ai-coding-assistant-07-context">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 13 Feb 2026 00:00:00 GMT</pubDate>
      <category>devtools</category>
      <category>ai-coding</category>
      <category>claude-code</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[6장: Elasticsearch AI 검색 통합]]></title>
      <link>https://archive.kreathlab.com/tech/ai-search-06-elasticsearch</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-search-06-elasticsearch</guid>
      <description><![CDATA[Elasticsearch의 kNN 검색, Inference API, semantic_text 필드, ELSER, Elastic Rerank, 하이브리드 검색(RRF)을 실습과 함께 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Elasticsearch의 kNN 검색, Inference API, semantic_text 필드, ELSER, Elastic Rerank, 하이브리드 검색(RRF)을 실습과 함께 다룹니다.</p><p><strong>AI / ML</strong> / 12분 읽기 / 시리즈: ai-search</p><p><a href="https://archive.kreathlab.com/tech/ai-search-06-elasticsearch">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 13 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>search</category>
      <category>ai</category>
    </item>
    <item>
      <title><![CDATA[4장: 문서 이해와 OCR 통합]]></title>
      <link>https://archive.kreathlab.com/tech/multimodal-ai-04-document-understanding</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/multimodal-ai-04-document-understanding</guid>
      <description><![CDATA[멀티모달 AI를 활용한 문서 이해 기법 — PDF 분석, 표 추출, 양식 처리, OCR 통합, 그리고 문서 처리 파이프라인 설계를 실전 중심으로 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>멀티모달 AI를 활용한 문서 이해 기법 — PDF 분석, 표 추출, 양식 처리, OCR 통합, 그리고 문서 처리 파이프라인 설계를 실전 중심으로 다룹니다.</p><p><strong>AI / ML</strong> / 12분 읽기 / 시리즈: multimodal-ai</p><p><a href="https://archive.kreathlab.com/tech/multimodal-ai-04-document-understanding">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 13 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>multimodal</category>
      <category>embedding</category>
    </item>
    <item>
      <title><![CDATA[13장: 실전 프로젝트 - Next.js 풀스택 앱 구축]]></title>
      <link>https://archive.kreathlab.com/tech/nextjs-app-router-13-practical-project</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/nextjs-app-router-13-practical-project</guid>
      <description><![CDATA[Link Shortener 앱을 구축하며 Next.js App Router의 모든 기능을 실전에 적용합니다. 프로젝트 구조 설계부터 배포까지 전 과정을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Link Shortener 앱을 구축하며 Next.js App Router의 모든 기능을 실전에 적용합니다. 프로젝트 구조 설계부터 배포까지 전 과정을 다룹니다.</p><p><strong>웹 개발</strong> / 21분 읽기 / 시리즈: nextjs-app-router</p><p><a href="https://archive.kreathlab.com/tech/nextjs-app-router-13-practical-project">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 13 Feb 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>nextjs</category>
      <category>react</category>
      <category>typescript</category>
      <category>frontend</category>
    </item>
    <item>
      <title><![CDATA[2장: 소유권과 차용 체크 멘탈 모델]]></title>
      <link>https://archive.kreathlab.com/tech/rust-backend-02-ownership</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/rust-backend-02-ownership</guid>
      <description><![CDATA[Rust의 핵심 개념인 소유권 규칙 3가지, Move 시맨틱스, 불변/가변 빌림, 라이프타임 기초를 백엔드 개발자 관점에서 체계적으로 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Rust의 핵심 개념인 소유권 규칙 3가지, Move 시맨틱스, 불변/가변 빌림, 라이프타임 기초를 백엔드 개발자 관점에서 체계적으로 다룹니다.</p><p><strong>프로그래밍</strong> / 17분 읽기 / 시리즈: rust-backend</p><p><a href="https://archive.kreathlab.com/tech/rust-backend-02-ownership">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 13 Feb 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>rust</category>
    </item>
    <item>
      <title><![CDATA[12장: 실전 프로젝트 — 타입 안전 유틸리티 라이브러리 설계]]></title>
      <link>https://archive.kreathlab.com/tech/typescript-deepdive-12-practical-project</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/typescript-deepdive-12-practical-project</guid>
      <description><![CDATA[TypeScript 5.x의 고급 타입 기법을 총동원하여 타입 안전 유틸리티 라이브러리를 처음부터 설계하고 구현하는 실전 프로젝트입니다.]]></description>
      <content:encoded><![CDATA[<p>TypeScript 5.x의 고급 타입 기법을 총동원하여 타입 안전 유틸리티 라이브러리를 처음부터 설계하고 구현하는 실전 프로젝트입니다.</p><p><strong>프로그래밍</strong> / 18분 읽기 / 시리즈: typescript-deepdive</p><p><a href="https://archive.kreathlab.com/tech/typescript-deepdive-12-practical-project">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 13 Feb 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>typescript</category>
      <category>performance</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[6장: Haystack -- 모듈러 파이프라인 아키텍처]]></title>
      <link>https://archive.kreathlab.com/tech/ai-orchestration-06-haystack</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-orchestration-06-haystack</guid>
      <description><![CDATA[deepset Haystack 2.x의 컴포넌트와 파이프라인 개념, 방향성 멀티그래프, AsyncPipeline, 라우터, 문서 스토어를 분석합니다.]]></description>
      <content:encoded><![CDATA[<p>deepset Haystack 2.x의 컴포넌트와 파이프라인 개념, 방향성 멀티그래프, AsyncPipeline, 라우터, 문서 스토어를 분석합니다.</p><p><strong>AI / ML</strong> / 14분 읽기 / 시리즈: ai-orchestration</p><p><a href="https://archive.kreathlab.com/tech/ai-orchestration-06-haystack">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 12 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>orchestration</category>
      <category>ai-framework</category>
      <category>ai</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[5장: AI 서비스 API 설계 패턴]]></title>
      <link>https://archive.kreathlab.com/tech/api-design-05-ai-patterns</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/api-design-05-ai-patterns</guid>
      <description><![CDATA[비동기 작업 패턴, 멀티모달 입력 처리, Function Calling 인터페이스, 배치 API, 구조화된 출력 등 AI 서비스 고유의 API 설계 패턴을 학습합니다.]]></description>
      <content:encoded><![CDATA[<p>비동기 작업 패턴, 멀티모달 입력 처리, Function Calling 인터페이스, 배치 API, 구조화된 출력 등 AI 서비스 고유의 API 설계 패턴을 학습합니다.</p><p><strong>아키텍처</strong> / 17분 읽기 / 시리즈: api-design</p><p><a href="https://archive.kreathlab.com/tech/api-design-05-ai-patterns">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 12 Feb 2026 00:00:00 GMT</pubDate>
      <category>architecture</category>
      <category>api-design</category>
      <category>graphql</category>
      <category>architecture</category>
    </item>
    <item>
      <title><![CDATA[4장: N+1 문제와 DataLoader 최적화]]></title>
      <link>https://archive.kreathlab.com/tech/graphql-architecture-04-n1-dataloader</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/graphql-architecture-04-n1-dataloader</guid>
      <description><![CDATA[GraphQL에서 가장 흔한 성능 문제인 N+1 문제의 원인을 분석하고, DataLoader를 통한 배칭과 캐싱 최적화를 구현합니다. 쿼리 복잡도 분석과 깊이 제한 전략도 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>GraphQL에서 가장 흔한 성능 문제인 N+1 문제의 원인을 분석하고, DataLoader를 통한 배칭과 캐싱 최적화를 구현합니다. 쿼리 복잡도 분석과 깊이 제한 전략도 다룹니다.</p><p><strong>웹 개발</strong> / 13분 읽기 / 시리즈: graphql-architecture</p><p><a href="https://archive.kreathlab.com/tech/graphql-architecture-04-n1-dataloader">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 12 Feb 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>graphql</category>
      <category>api-design</category>
      <category>frontend</category>
    </item>
    <item>
      <title><![CDATA[3장: 분산 추적(Distributed Tracing)]]></title>
      <link>https://archive.kreathlab.com/tech/opentelemetry-03-traces</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/opentelemetry-03-traces</guid>
      <description><![CDATA[스팬의 내부 구조와 종류, 부모-자식 관계, 샘플링 전략(Head/Tail/Rate)을 학습하고 Python으로 분산 추적을 직접 구현합니다.]]></description>
      <content:encoded><![CDATA[<p>스팬의 내부 구조와 종류, 부모-자식 관계, 샘플링 전략(Head/Tail/Rate)을 학습하고 Python으로 분산 추적을 직접 구현합니다.</p><p><strong>인프라</strong> / 15분 읽기 / 시리즈: opentelemetry</p><p><a href="https://archive.kreathlab.com/tech/opentelemetry-03-traces">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 12 Feb 2026 00:00:00 GMT</pubDate>
      <category>infra</category>
      <category>monitoring</category>
      <category>observability</category>
    </item>
    <item>
      <title><![CDATA[11장: 실전 프로젝트 - React 19 풀스택 앱 구축]]></title>
      <link>https://archive.kreathlab.com/tech/react19-rsc-11-practical-project</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/react19-rsc-11-practical-project</guid>
      <description><![CDATA[React 19의 핵심 기능을 모두 활용한 풀스택 북마크 앱을 구축합니다. Server Components, Server Actions, 새로운 훅, Suspense 패턴을 실전에 적용합니다.]]></description>
      <content:encoded><![CDATA[<p>React 19의 핵심 기능을 모두 활용한 풀스택 북마크 앱을 구축합니다. Server Components, Server Actions, 새로운 훅, Suspense 패턴을 실전에 적용합니다.</p><p><strong>웹 개발</strong> / 16분 읽기 / 시리즈: react19-rsc</p><p><a href="https://archive.kreathlab.com/tech/react19-rsc-11-practical-project">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 12 Feb 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>react</category>
      <category>nextjs</category>
      <category>performance</category>
      <category>frontend</category>
      <category>typescript</category>
    </item>
    <item>
      <title><![CDATA[2장: 이벤트 스트리밍 아키텍처 기초]]></title>
      <link>https://archive.kreathlab.com/tech/realtime-pipeline-02-event-streaming</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/realtime-pipeline-02-event-streaming</guid>
      <description><![CDATA[이벤트 로그, 토픽과 파티션, 오프셋 관리, 이벤트 시간과 처리 시간의 차이, 워터마크, 윈도우 연산 등 스트림 처리의 핵심 개념을 체계적으로 학습합니다.]]></description>
      <content:encoded><![CDATA[<p>이벤트 로그, 토픽과 파티션, 오프셋 관리, 이벤트 시간과 처리 시간의 차이, 워터마크, 윈도우 연산 등 스트림 처리의 핵심 개념을 체계적으로 학습합니다.</p><p><strong>데이터</strong> / 20분 읽기 / 시리즈: realtime-pipeline</p><p><a href="https://archive.kreathlab.com/tech/realtime-pipeline-02-event-streaming">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 12 Feb 2026 00:00:00 GMT</pubDate>
      <category>data</category>
      <category>streaming</category>
      <category>data-engineering</category>
    </item>
    <item>
      <title><![CDATA[6장: Windsurf, Codex, 기타 도구들]]></title>
      <link>https://archive.kreathlab.com/tech/ai-coding-assistant-06-others</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-coding-assistant-06-others</guid>
      <description><![CDATA[Windsurf Cascade, OpenAI Codex CLI, Google Antigravity, Amazon Q, Gemini Code Assist, JetBrains AI, Aider 등 주요 AI 코딩 도구들의 특성과 포지셔닝을 비교합니다.]]></description>
      <content:encoded><![CDATA[<p>Windsurf Cascade, OpenAI Codex CLI, Google Antigravity, Amazon Q, Gemini Code Assist, JetBrains AI, Aider 등 주요 AI 코딩 도구들의 특성과 포지셔닝을 비교합니다.</p><p><strong>개발 도구</strong> / 17분 읽기 / 시리즈: ai-coding-assistant</p><p><a href="https://archive.kreathlab.com/tech/ai-coding-assistant-06-others">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 11 Feb 2026 00:00:00 GMT</pubDate>
      <category>devtools</category>
      <category>ai-coding</category>
      <category>claude-code</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[5장: 랭킹 모델 — Bi-encoder와 Cross-encoder]]></title>
      <link>https://archive.kreathlab.com/tech/ai-search-05-ranking</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-search-05-ranking</guid>
      <description><![CDATA[Bi-encoder와 Cross-encoder의 구조적 차이, Elastic Rerank의 DeBERTa v3 모델, 점수 퓨전(RRF), 학습 랭킹(LTR)을 심층적으로 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Bi-encoder와 Cross-encoder의 구조적 차이, Elastic Rerank의 DeBERTa v3 모델, 점수 퓨전(RRF), 학습 랭킹(LTR)을 심층적으로 다룹니다.</p><p><strong>AI / ML</strong> / 14분 읽기 / 시리즈: ai-search</p><p><a href="https://archive.kreathlab.com/tech/ai-search-05-ranking">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 11 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>search</category>
      <category>ai</category>
    </item>
    <item>
      <title><![CDATA[3장: 이미지 이해와 시각적 질의응답]]></title>
      <link>https://archive.kreathlab.com/tech/multimodal-ai-03-image-understanding</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/multimodal-ai-03-image-understanding</guid>
      <description><![CDATA[멀티모달 AI를 활용한 이미지 이해의 실전 기법 — 시각적 질의응답, 이미지 분석 프롬프트 설계, 정확도 향상 전략, 그리고 다양한 활용 패턴을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>멀티모달 AI를 활용한 이미지 이해의 실전 기법 — 시각적 질의응답, 이미지 분석 프롬프트 설계, 정확도 향상 전략, 그리고 다양한 활용 패턴을 다룹니다.</p><p><strong>AI / ML</strong> / 13분 읽기 / 시리즈: multimodal-ai</p><p><a href="https://archive.kreathlab.com/tech/multimodal-ai-03-image-understanding">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 11 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>multimodal</category>
      <category>embedding</category>
    </item>
    <item>
      <title><![CDATA[12장: 모니터링, 보안, 프로덕션 배포]]></title>
      <link>https://archive.kreathlab.com/tech/nextjs-app-router-12-deployment</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/nextjs-app-router-12-deployment</guid>
      <description><![CDATA[Next.js의 프로덕션 배포를 다룹니다. instrumentation.ts, OpenTelemetry, 보안 헤더, Docker 배포, Build Adapters, 셀프 호스팅 전략을 살펴봅니다.]]></description>
      <content:encoded><![CDATA[<p>Next.js의 프로덕션 배포를 다룹니다. instrumentation.ts, OpenTelemetry, 보안 헤더, Docker 배포, Build Adapters, 셀프 호스팅 전략을 살펴봅니다.</p><p><strong>웹 개발</strong> / 15분 읽기 / 시리즈: nextjs-app-router</p><p><a href="https://archive.kreathlab.com/tech/nextjs-app-router-12-deployment">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 11 Feb 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>nextjs</category>
      <category>react</category>
      <category>typescript</category>
      <category>frontend</category>
    </item>
    <item>
      <title><![CDATA[1장: 백엔드 개발자가 Rust를 배워야 하는 이유]]></title>
      <link>https://archive.kreathlab.com/tech/rust-backend-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/rust-backend-01-introduction</guid>
      <description><![CDATA[Rust의 고유 가치인 안전성과 성능, 백엔드 생태계 현황, GC 없는 메모리 관리, 2026년 채택 현황까지 백엔드 개발자 관점에서 Rust를 배워야 하는 이유를 살펴봅니다.]]></description>
      <content:encoded><![CDATA[<p>Rust의 고유 가치인 안전성과 성능, 백엔드 생태계 현황, GC 없는 메모리 관리, 2026년 채택 현황까지 백엔드 개발자 관점에서 Rust를 배워야 하는 이유를 살펴봅니다.</p><p><strong>프로그래밍</strong> / 15분 읽기 / 시리즈: rust-backend</p><p><a href="https://archive.kreathlab.com/tech/rust-backend-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 11 Feb 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>rust</category>
    </item>
    <item>
      <title><![CDATA[11장: 프로젝트 설정과 모노레포 타입 전략]]></title>
      <link>https://archive.kreathlab.com/tech/typescript-deepdive-11-project-monorepo</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/typescript-deepdive-11-project-monorepo</guid>
      <description><![CDATA[TypeScript 프로젝트의 tsconfig.json 최적화, 프로젝트 참조, Isolated Declarations, 모노레포에서의 타입 전략을 실전 중심으로 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>TypeScript 프로젝트의 tsconfig.json 최적화, 프로젝트 참조, Isolated Declarations, 모노레포에서의 타입 전략을 실전 중심으로 다룹니다.</p><p><strong>프로그래밍</strong> / 14분 읽기 / 시리즈: typescript-deepdive</p><p><a href="https://archive.kreathlab.com/tech/typescript-deepdive-11-project-monorepo">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 11 Feb 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>typescript</category>
      <category>performance</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[5장: Semantic Kernel -- 엔터프라이즈 AI 오케스트레이션]]></title>
      <link>https://archive.kreathlab.com/tech/ai-orchestration-05-semantic-kernel</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-orchestration-05-semantic-kernel</guid>
      <description><![CDATA[Microsoft Semantic Kernel의 멀티 언어 아키텍처, 플러그인 시스템, 플래너, Azure 통합, 엔터프라이즈 보안과 거버넌스를 분석합니다.]]></description>
      <content:encoded><![CDATA[<p>Microsoft Semantic Kernel의 멀티 언어 아키텍처, 플러그인 시스템, 플래너, Azure 통합, 엔터프라이즈 보안과 거버넌스를 분석합니다.</p><p><strong>AI / ML</strong> / 14분 읽기 / 시리즈: ai-orchestration</p><p><a href="https://archive.kreathlab.com/tech/ai-orchestration-05-semantic-kernel">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 10 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>orchestration</category>
      <category>ai-framework</category>
      <category>ai</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[4장: GraphQL — 유연한 데이터 쿼리]]></title>
      <link>https://archive.kreathlab.com/tech/api-design-04-graphql</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/api-design-04-graphql</guid>
      <description><![CDATA[GraphQL의 스키마 퍼스트 설계, 타입 시스템, N+1 문제 해결, AI 서비스 데이터 모델링을 Apollo Server 실습과 함께 학습합니다.]]></description>
      <content:encoded><![CDATA[<p>GraphQL의 스키마 퍼스트 설계, 타입 시스템, N+1 문제 해결, AI 서비스 데이터 모델링을 Apollo Server 실습과 함께 학습합니다.</p><p><strong>아키텍처</strong> / 12분 읽기 / 시리즈: api-design</p><p><a href="https://archive.kreathlab.com/tech/api-design-04-graphql">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 10 Feb 2026 00:00:00 GMT</pubDate>
      <category>architecture</category>
      <category>api-design</category>
      <category>graphql</category>
      <category>architecture</category>
    </item>
    <item>
      <title><![CDATA[3장: Resolver 구현과 데이터 로딩]]></title>
      <link>https://archive.kreathlab.com/tech/graphql-architecture-03-resolvers</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/graphql-architecture-03-resolvers</guid>
      <description><![CDATA[GraphQL 리졸버 체인의 실행 흐름과 컨텍스트 설계를 이해합니다. 데이터 소스 추상화, 부모-자식 리졸버 관계, Apollo Server와 GraphQL Yoga에서의 실전 구현 패턴을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>GraphQL 리졸버 체인의 실행 흐름과 컨텍스트 설계를 이해합니다. 데이터 소스 추상화, 부모-자식 리졸버 관계, Apollo Server와 GraphQL Yoga에서의 실전 구현 패턴을 다룹니다.</p><p><strong>웹 개발</strong> / 14분 읽기 / 시리즈: graphql-architecture</p><p><a href="https://archive.kreathlab.com/tech/graphql-architecture-03-resolvers">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 10 Feb 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>graphql</category>
      <category>api-design</category>
      <category>frontend</category>
    </item>
    <item>
      <title><![CDATA[2장: OpenTelemetry 아키텍처 심층 분석]]></title>
      <link>https://archive.kreathlab.com/tech/opentelemetry-02-architecture</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/opentelemetry-02-architecture</guid>
      <description><![CDATA[OpenTelemetry의 API/SDK/Collector 3계층 구조, W3C TraceContext 기반 컨텍스트 전파, 리소스와 시맨틱 컨벤션, 배포 패턴을 심층적으로 분석합니다.]]></description>
      <content:encoded><![CDATA[<p>OpenTelemetry의 API/SDK/Collector 3계층 구조, W3C TraceContext 기반 컨텍스트 전파, 리소스와 시맨틱 컨벤션, 배포 패턴을 심층적으로 분석합니다.</p><p><strong>인프라</strong> / 16분 읽기 / 시리즈: opentelemetry</p><p><a href="https://archive.kreathlab.com/tech/opentelemetry-02-architecture">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 10 Feb 2026 00:00:00 GMT</pubDate>
      <category>infra</category>
      <category>monitoring</category>
      <category>observability</category>
    </item>
    <item>
      <title><![CDATA[10장: React 18에서 19로 마이그레이션]]></title>
      <link>https://archive.kreathlab.com/tech/react19-rsc-10-migration</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/react19-rsc-10-migration</guid>
      <description><![CDATA[React 18에서 19로 안전하게 업그레이드하는 단계별 가이드입니다. 제거된 API, 타입 변경, 동작 변화, 자동 마이그레이션 도구를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>React 18에서 19로 안전하게 업그레이드하는 단계별 가이드입니다. 제거된 API, 타입 변경, 동작 변화, 자동 마이그레이션 도구를 다룹니다.</p><p><strong>웹 개발</strong> / 14분 읽기 / 시리즈: react19-rsc</p><p><a href="https://archive.kreathlab.com/tech/react19-rsc-10-migration">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 10 Feb 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>react</category>
      <category>nextjs</category>
      <category>performance</category>
      <category>frontend</category>
      <category>typescript</category>
    </item>
    <item>
      <title><![CDATA[1장: 실시간 데이터 파이프라인의 필요성과 핵심 개념]]></title>
      <link>https://archive.kreathlab.com/tech/realtime-pipeline-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/realtime-pipeline-01-introduction</guid>
      <description><![CDATA[배치와 실시간 처리의 차이, 이벤트 드리븐 아키텍처, Lambda/Kappa 아키텍처, 핵심 구성요소를 살펴보며 실시간 데이터 파이프라인의 전체 그림을 이해합니다.]]></description>
      <content:encoded><![CDATA[<p>배치와 실시간 처리의 차이, 이벤트 드리븐 아키텍처, Lambda/Kappa 아키텍처, 핵심 구성요소를 살펴보며 실시간 데이터 파이프라인의 전체 그림을 이해합니다.</p><p><strong>데이터</strong> / 18분 읽기 / 시리즈: realtime-pipeline</p><p><a href="https://archive.kreathlab.com/tech/realtime-pipeline-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 10 Feb 2026 00:00:00 GMT</pubDate>
      <category>data</category>
      <category>streaming</category>
      <category>data-engineering</category>
    </item>
    <item>
      <title><![CDATA[5장: Claude Code — CLI 기반 에이전틱 코딩]]></title>
      <link>https://archive.kreathlab.com/tech/ai-coding-assistant-05-claude-code</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-coding-assistant-05-claude-code</guid>
      <description><![CDATA[Claude Code의 터미널 에이전트 아키텍처, CLAUDE.md 컨텍스트, 도구 시스템, 서브에이전트, Git 워크플로우 통합, Hooks, MCP 서버 연동을 분석합니다.]]></description>
      <content:encoded><![CDATA[<p>Claude Code의 터미널 에이전트 아키텍처, CLAUDE.md 컨텍스트, 도구 시스템, 서브에이전트, Git 워크플로우 통합, Hooks, MCP 서버 연동을 분석합니다.</p><p><strong>개발 도구</strong> / 16분 읽기 / 시리즈: ai-coding-assistant</p><p><a href="https://archive.kreathlab.com/tech/ai-coding-assistant-05-claude-code">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 09 Feb 2026 00:00:00 GMT</pubDate>
      <category>devtools</category>
      <category>ai-coding</category>
      <category>claude-code</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[4장: 쿼리 이해와 확장]]></title>
      <link>https://archive.kreathlab.com/tech/ai-search-04-query-understanding</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-search-04-query-understanding</guid>
      <description><![CDATA[쿼리 분류, 의도 인식, 엔티티 인식부터 LLM 기반 쿼리 확장, HyDE(가상 문서 생성), 다국어 처리까지 쿼리 이해 파이프라인을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>쿼리 분류, 의도 인식, 엔티티 인식부터 LLM 기반 쿼리 확장, HyDE(가상 문서 생성), 다국어 처리까지 쿼리 이해 파이프라인을 다룹니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: ai-search</p><p><a href="https://archive.kreathlab.com/tech/ai-search-04-query-understanding">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 09 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>search</category>
      <category>ai</category>
    </item>
    <item>
      <title><![CDATA[2장: Vision-Language 모델 아키텍처]]></title>
      <link>https://archive.kreathlab.com/tech/multimodal-ai-02-vlm-architecture</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/multimodal-ai-02-vlm-architecture</guid>
      <description><![CDATA[CLIP에서 GPT-4o까지, Vision-Language 모델의 핵심 아키텍처를 분석합니다. 이미지 인코딩, 크로스모달 정렬, 통합 모델의 설계 원리를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>CLIP에서 GPT-4o까지, Vision-Language 모델의 핵심 아키텍처를 분석합니다. 이미지 인코딩, 크로스모달 정렬, 통합 모델의 설계 원리를 다룹니다.</p><p><strong>AI / ML</strong> / 12분 읽기 / 시리즈: multimodal-ai</p><p><a href="https://archive.kreathlab.com/tech/multimodal-ai-02-vlm-architecture">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 09 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>multimodal</category>
      <category>embedding</category>
    </item>
    <item>
      <title><![CDATA[11장: 이미지, 폰트, 메타데이터 최적화]]></title>
      <link>https://archive.kreathlab.com/tech/nextjs-app-router-11-optimization</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/nextjs-app-router-11-optimization</guid>
      <description><![CDATA[Next.js의 이미지, 폰트, 메타데이터 최적화를 다룹니다. next/image, next/font, generateMetadata, OG 이미지 생성, Core Web Vitals 전략을 살펴봅니다.]]></description>
      <content:encoded><![CDATA[<p>Next.js의 이미지, 폰트, 메타데이터 최적화를 다룹니다. next/image, next/font, generateMetadata, OG 이미지 생성, Core Web Vitals 전략을 살펴봅니다.</p><p><strong>웹 개발</strong> / 17분 읽기 / 시리즈: nextjs-app-router</p><p><a href="https://archive.kreathlab.com/tech/nextjs-app-router-11-optimization">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 09 Feb 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>nextjs</category>
      <category>react</category>
      <category>typescript</category>
      <category>frontend</category>
    </item>
    <item>
      <title><![CDATA[13장: Python 3.13 마이그레이션 실전 가이드]]></title>
      <link>https://archive.kreathlab.com/tech/python-trends-13-migration-guide</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/python-trends-13-migration-guide</guid>
      <description><![CDATA[기존 프로젝트를 Python 3.13으로 업그레이드하는 실전 가이드입니다. 호환성 체크리스트, 단계별 전략, 주요 라이브러리 호환성, 도구 전환 계획을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>기존 프로젝트를 Python 3.13으로 업그레이드하는 실전 가이드입니다. 호환성 체크리스트, 단계별 전략, 주요 라이브러리 호환성, 도구 전환 계획을 다룹니다.</p><p><strong>프로그래밍</strong> / 17분 읽기 / 시리즈: python-trends</p><p><a href="https://archive.kreathlab.com/tech/python-trends-13-migration-guide">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 09 Feb 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>python</category>
      <category>typescript</category>
      <category>performance</category>
      <category>devtools</category>
      <category>concurrency</category>
    </item>
    <item>
      <title><![CDATA[10장: 타입 수준 프로그래밍 — 타입으로 로직 작성하기]]></title>
      <link>https://archive.kreathlab.com/tech/typescript-deepdive-10-type-level-programming</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/typescript-deepdive-10-type-level-programming</guid>
      <description><![CDATA[TypeScript 타입 시스템을 프로그래밍 언어로 활용하는 고급 기법 — 산술 연산, 문자열 파서, 상태 머신 등을 타입만으로 구현하는 패턴을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>TypeScript 타입 시스템을 프로그래밍 언어로 활용하는 고급 기법 — 산술 연산, 문자열 파서, 상태 머신 등을 타입만으로 구현하는 패턴을 다룹니다.</p><p><strong>프로그래밍</strong> / 14분 읽기 / 시리즈: typescript-deepdive</p><p><a href="https://archive.kreathlab.com/tech/typescript-deepdive-10-type-level-programming">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 09 Feb 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>typescript</category>
      <category>performance</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[4장: LlamaIndex -- 데이터 프레임워크와 워크플로우]]></title>
      <link>https://archive.kreathlab.com/tech/ai-orchestration-04-llamaindex</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-orchestration-04-llamaindex</guid>
      <description><![CDATA[LlamaIndex의 데이터 커넥터, 인덱스 유형, 쿼리 엔진, 그리고 이벤트 드리븐 Workflows 1.0을 실전 예제와 함께 분석합니다.]]></description>
      <content:encoded><![CDATA[<p>LlamaIndex의 데이터 커넥터, 인덱스 유형, 쿼리 엔진, 그리고 이벤트 드리븐 Workflows 1.0을 실전 예제와 함께 분석합니다.</p><p><strong>AI / ML</strong> / 15분 읽기 / 시리즈: ai-orchestration</p><p><a href="https://archive.kreathlab.com/tech/ai-orchestration-04-llamaindex">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 08 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>orchestration</category>
      <category>ai-framework</category>
      <category>ai</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[3장: gRPC — 고성능 서비스 간 통신]]></title>
      <link>https://archive.kreathlab.com/tech/api-design-03-grpc</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/api-design-03-grpc</guid>
      <description><![CDATA[HTTP/2와 Protocol Buffers 기반의 gRPC를 활용한 고성능 마이크로서비스 통신을 학습합니다. 4가지 스트리밍 모드와 AI 추론 서비스 구현을 실습합니다.]]></description>
      <content:encoded><![CDATA[<p>HTTP/2와 Protocol Buffers 기반의 gRPC를 활용한 고성능 마이크로서비스 통신을 학습합니다. 4가지 스트리밍 모드와 AI 추론 서비스 구현을 실습합니다.</p><p><strong>아키텍처</strong> / 15분 읽기 / 시리즈: api-design</p><p><a href="https://archive.kreathlab.com/tech/api-design-03-grpc">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 08 Feb 2026 00:00:00 GMT</pubDate>
      <category>architecture</category>
      <category>api-design</category>
      <category>graphql</category>
      <category>architecture</category>
    </item>
    <item>
      <title><![CDATA[2장: 스키마 설계 패턴]]></title>
      <link>https://archive.kreathlab.com/tech/graphql-architecture-02-schema-design</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/graphql-architecture-02-schema-design</guid>
      <description><![CDATA[GraphQL 스키마 설계의 두 가지 접근법과 실전 타입 설계 패턴을 다룹니다. Node 인터페이스, Relay 커서 페이지네이션, 유니온/인터페이스 등 프로덕션 수준의 스키마 설계 원칙을 학습합니다.]]></description>
      <content:encoded><![CDATA[<p>GraphQL 스키마 설계의 두 가지 접근법과 실전 타입 설계 패턴을 다룹니다. Node 인터페이스, Relay 커서 페이지네이션, 유니온/인터페이스 등 프로덕션 수준의 스키마 설계 원칙을 학습합니다.</p><p><strong>웹 개발</strong> / 13분 읽기 / 시리즈: graphql-architecture</p><p><a href="https://archive.kreathlab.com/tech/graphql-architecture-02-schema-design">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 08 Feb 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>graphql</category>
      <category>api-design</category>
      <category>frontend</category>
    </item>
    <item>
      <title><![CDATA[10장: 실전 프로젝트 - 풀스택 MCP 시스템 구축]]></title>
      <link>https://archive.kreathlab.com/tech/mcp-guide-10-fullstack-project</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/mcp-guide-10-fullstack-project</guid>
      <description><![CDATA[여러 MCP 서버를 조합하고 클라이언트 애플리케이션을 구축하여, 프로덕션 수준의 풀스택 MCP 시스템을 완성하는 실전 프로젝트입니다.]]></description>
      <content:encoded><![CDATA[<p>여러 MCP 서버를 조합하고 클라이언트 애플리케이션을 구축하여, 프로덕션 수준의 풀스택 MCP 시스템을 완성하는 실전 프로젝트입니다.</p><p><strong>AI / ML</strong> / 22분 읽기 / 시리즈: mcp-guide</p><p><a href="https://archive.kreathlab.com/tech/mcp-guide-10-fullstack-project">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 08 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>mcp</category>
      <category>typescript</category>
      <category>python</category>
    </item>
    <item>
      <title><![CDATA[1장: 관측 가능성의 진화와 OpenTelemetry]]></title>
      <link>https://archive.kreathlab.com/tech/opentelemetry-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/opentelemetry-01-introduction</guid>
      <description><![CDATA[로깅에서 APM, 관측 가능성으로 이어지는 모니터링의 진화를 살펴보고, OpenTelemetry가 탄생한 배경과 3대 신호, 벤더 중립의 가치를 이해합니다.]]></description>
      <content:encoded><![CDATA[<p>로깅에서 APM, 관측 가능성으로 이어지는 모니터링의 진화를 살펴보고, OpenTelemetry가 탄생한 배경과 3대 신호, 벤더 중립의 가치를 이해합니다.</p><p><strong>인프라</strong> / 18분 읽기 / 시리즈: opentelemetry</p><p><a href="https://archive.kreathlab.com/tech/opentelemetry-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 08 Feb 2026 00:00:00 GMT</pubDate>
      <category>infra</category>
      <category>monitoring</category>
      <category>observability</category>
    </item>
    <item>
      <title><![CDATA[9장: 성능 최적화 전략과 베스트 프랙티스]]></title>
      <link>https://archive.kreathlab.com/tech/react19-rsc-09-performance</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/react19-rsc-09-performance</guid>
      <description><![CDATA[React 19 애플리케이션의 성능을 극대화하는 전략을 다룹니다. 번들 최적화, 렌더링 성능, Core Web Vitals 개선, 측정 도구 활용법을 배웁���다.]]></description>
      <content:encoded><![CDATA[<p>React 19 애플리케이션의 성능을 극대화하는 전략을 다룹니다. 번들 최적화, 렌더링 성능, Core Web Vitals 개선, 측정 도구 활용법을 배웁���다.</p><p><strong>웹 개발</strong> / 15분 읽기 / 시리즈: react19-rsc</p><p><a href="https://archive.kreathlab.com/tech/react19-rsc-09-performance">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 08 Feb 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>react</category>
      <category>nextjs</category>
      <category>performance</category>
      <category>frontend</category>
      <category>typescript</category>
    </item>
    <item>
      <title><![CDATA[4장: Cursor — AI-native IDE의 선두주자]]></title>
      <link>https://archive.kreathlab.com/tech/ai-coding-assistant-04-cursor</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-coding-assistant-04-cursor</guid>
      <description><![CDATA[Cursor의 아키텍처, Composer 멀티파일 편집, Background Agents, Tab 자동완성, .cursorrules, 자기 요약 기법까지 심층 분석합니다.]]></description>
      <content:encoded><![CDATA[<p>Cursor의 아키텍처, Composer 멀티파일 편집, Background Agents, Tab 자동완성, .cursorrules, 자기 요약 기법까지 심층 분석합니다.</p><p><strong>개발 도구</strong> / 16분 읽기 / 시리즈: ai-coding-assistant</p><p><a href="https://archive.kreathlab.com/tech/ai-coding-assistant-04-cursor">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 07 Feb 2026 00:00:00 GMT</pubDate>
      <category>devtools</category>
      <category>ai-coding</category>
      <category>claude-code</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[3장: 검색 품질 메트릭과 평가]]></title>
      <link>https://archive.kreathlab.com/tech/ai-search-03-metrics</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-search-03-metrics</guid>
      <description><![CDATA[Precision, Recall, NDCG, MRR, MAP 등 검색 품질 메트릭의 원리와 계산법, 오프라인/온라인 평가 방법론, A/B 테스트와 평가 데이터셋 구축을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Precision, Recall, NDCG, MRR, MAP 등 검색 품질 메트릭의 원리와 계산법, 오프라인/온라인 평가 방법론, A/B 테스트와 평가 데이터셋 구축을 다룹니다.</p><p><strong>AI / ML</strong> / 15분 읽기 / 시리즈: ai-search</p><p><a href="https://archive.kreathlab.com/tech/ai-search-03-metrics">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 07 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>search</category>
      <category>ai</category>
    </item>
    <item>
      <title><![CDATA[1장: 멀티모달 AI의 등장과 핵심 개념]]></title>
      <link>https://archive.kreathlab.com/tech/multimodal-ai-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/multimodal-ai-01-introduction</guid>
      <description><![CDATA[멀티모달 AI의 정의, 발전 역사, 핵심 아키텍처 패턴, 그리고 주요 모델(GPT-4o, Claude, Gemini)의 멀티모달 능력을 조망합니다.]]></description>
      <content:encoded><![CDATA[<p>멀티모달 AI의 정의, 발전 역사, 핵심 아키텍처 패턴, 그리고 주요 모델(GPT-4o, Claude, Gemini)의 멀티모달 능력을 조망합니다.</p><p><strong>AI / ML</strong> / 12분 읽기 / 시리즈: multimodal-ai</p><p><a href="https://archive.kreathlab.com/tech/multimodal-ai-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 07 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>multimodal</category>
      <category>embedding</category>
    </item>
    <item>
      <title><![CDATA[10장: 미들웨어와 Proxy 고급 패턴]]></title>
      <link>https://archive.kreathlab.com/tech/nextjs-app-router-10-middleware</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/nextjs-app-router-10-middleware</guid>
      <description><![CDATA[Next.js의 미들웨어 진화를 다룹니다. middleware.ts에서 proxy.ts로의 전환, 인증, i18n, A/B 테스팅, 속도 제한 등 고급 패턴을 살펴봅니다.]]></description>
      <content:encoded><![CDATA[<p>Next.js의 미들웨어 진화를 다룹니다. middleware.ts에서 proxy.ts로의 전환, 인증, i18n, A/B 테스팅, 속도 제한 등 고급 패턴을 살펴봅니다.</p><p><strong>웹 개발</strong> / 17분 읽기 / 시리즈: nextjs-app-router</p><p><a href="https://archive.kreathlab.com/tech/nextjs-app-router-10-middleware">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 07 Feb 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>nextjs</category>
      <category>react</category>
      <category>typescript</category>
      <category>frontend</category>
    </item>
    <item>
      <title><![CDATA[12장: AI 개발에서의 Python 활용]]></title>
      <link>https://archive.kreathlab.com/tech/python-trends-12-ai-python</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/python-trends-12-ai-python</guid>
      <description><![CDATA[AI/ML 개발에서 Python이 차지하는 위치와 최신 트렌드를 다룹니다. PyTorch 생태계, LLM 개발 도구, 타입 안전한 AI 파이프라인, free-threaded Python의 AI 활용을 살펴봅니다.]]></description>
      <content:encoded><![CDATA[<p>AI/ML 개발에서 Python이 차지하는 위치와 최신 트렌드를 다룹니다. PyTorch 생태계, LLM 개발 도구, 타입 안전한 AI 파이프라인, free-threaded Python의 AI 활용을 살펴봅니다.</p><p><strong>프로그래밍</strong> / 13분 읽기 / 시리즈: python-trends</p><p><a href="https://archive.kreathlab.com/tech/python-trends-12-ai-python">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 07 Feb 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>python</category>
      <category>typescript</category>
      <category>performance</category>
      <category>devtools</category>
      <category>concurrency</category>
    </item>
    <item>
      <title><![CDATA[9장: NoInfer와 새로운 유틸리티 타입]]></title>
      <link>https://archive.kreathlab.com/tech/typescript-deepdive-09-noinfer-utilities</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/typescript-deepdive-09-noinfer-utilities</guid>
      <description><![CDATA[TypeScript 5.4의 NoInfer 유틸리티 타입과 5.x에서 추가된 새로운 타입 도구들을 활용한 라이브러리 설계 패턴을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>TypeScript 5.4의 NoInfer 유틸리티 타입과 5.x에서 추가된 새로운 타입 도구들을 활용한 라이브러리 설계 패턴을 다룹니다.</p><p><strong>프로그래밍</strong> / 14분 읽기 / 시리즈: typescript-deepdive</p><p><a href="https://archive.kreathlab.com/tech/typescript-deepdive-09-noinfer-utilities">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 07 Feb 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>typescript</category>
      <category>performance</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[10장: 실전 프로젝트 - AI 개발 워크플로우 통합 시스템]]></title>
      <link>https://archive.kreathlab.com/tech/ai-dev-workflow-10-practical-project</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-dev-workflow-10-practical-project</guid>
      <description><![CDATA[전체 시리즈에서 다룬 AI 코드 리뷰, 테스트 생성, 문서화, PR 분석을 하나의 통합 시스템으로 구축하는 실전 프로젝트를 진행합니다.]]></description>
      <content:encoded><![CDATA[<p>전체 시리즈에서 다룬 AI 코드 리뷰, 테스트 생성, 문서화, PR 분석을 하나의 통합 시스템으로 구축하는 실전 프로젝트를 진행합니다.</p><p><strong>AI / ML</strong> / 21분 읽기 / 시리즈: ai-dev-workflow</p><p><a href="https://archive.kreathlab.com/tech/ai-dev-workflow-10-practical-project">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 06 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>devtools</category>
      <category>automation</category>
      <category>code-quality</category>
      <category>devops</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[3장: LangGraph -- 그래프 기반 에이전트 오케스트레이션]]></title>
      <link>https://archive.kreathlab.com/tech/ai-orchestration-03-langgraph</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-orchestration-03-langgraph</guid>
      <description><![CDATA[LangGraph 1.0/1.1의 StateGraph, 듀러블 상태, 조건부 엣지, 휴먼인더루프, type-safe 스트리밍을 실전 예제와 함께 분석합니다.]]></description>
      <content:encoded><![CDATA[<p>LangGraph 1.0/1.1의 StateGraph, 듀러블 상태, 조건부 엣지, 휴먼인더루프, type-safe 스트리밍을 실전 예제와 함께 분석합니다.</p><p><strong>AI / ML</strong> / 14분 읽기 / 시리즈: ai-orchestration</p><p><a href="https://archive.kreathlab.com/tech/ai-orchestration-03-langgraph">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 06 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>orchestration</category>
      <category>ai-framework</category>
      <category>ai</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[2장: RESTful API 설계 원칙과 AI 서비스 적용]]></title>
      <link>https://archive.kreathlab.com/tech/api-design-02-rest</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/api-design-02-rest</guid>
      <description><![CDATA[Richardson 성숙도 모델부터 리소스 설계, HTTP 메서드, OpenAPI 3.1 스펙, AI 서비스 REST 엔드포인트 설계까지 RESTful API의 핵심 원칙을 실습합니다.]]></description>
      <content:encoded><![CDATA[<p>Richardson 성숙도 모델부터 리소스 설계, HTTP 메서드, OpenAPI 3.1 스펙, AI 서비스 REST 엔드포인트 설계까지 RESTful API의 핵심 원칙을 실습합니다.</p><p><strong>아키텍처</strong> / 16분 읽기 / 시리즈: api-design</p><p><a href="https://archive.kreathlab.com/tech/api-design-02-rest">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 06 Feb 2026 00:00:00 GMT</pubDate>
      <category>architecture</category>
      <category>api-design</category>
      <category>graphql</category>
      <category>architecture</category>
    </item>
    <item>
      <title><![CDATA[1장: GraphQL의 등장과 핵심 가치]]></title>
      <link>https://archive.kreathlab.com/tech/graphql-architecture-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/graphql-architecture-01-introduction</guid>
      <description><![CDATA[REST API의 한계인 오버페칭과 언더페칭 문제를 분석하고, GraphQL의 타입 시스템, 스키마, 리졸버 등 핵심 개념을 소개합니다. 2026년 GraphQL 생태계 현황과 도입 판단 기준을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>REST API의 한계인 오버페칭과 언더페칭 문제를 분석하고, GraphQL의 타입 시스템, 스키마, 리졸버 등 핵심 개념을 소개합니다. 2026년 GraphQL 생태계 현황과 도입 판단 기준을 다룹니다.</p><p><strong>웹 개발</strong> / 14분 읽기 / 시리즈: graphql-architecture</p><p><a href="https://archive.kreathlab.com/tech/graphql-architecture-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 06 Feb 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>graphql</category>
      <category>api-design</category>
      <category>frontend</category>
    </item>
    <item>
      <title><![CDATA[9장: 보안, 인증, 권한 관리]]></title>
      <link>https://archive.kreathlab.com/tech/mcp-guide-09-security</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/mcp-guide-09-security</guid>
      <description><![CDATA[MCP 서버의 OAuth 2.1 인증, 입력 유효성 검사, 명령 주입 방지, 데이터 유출 방지 등 프로덕션 보안 모범 사례를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>MCP 서버의 OAuth 2.1 인증, 입력 유효성 검사, 명령 주입 방지, 데이터 유출 방지 등 프로덕션 보안 모범 사례를 다룹니다.</p><p><strong>AI / ML</strong> / 20분 읽기 / 시리즈: mcp-guide</p><p><a href="https://archive.kreathlab.com/tech/mcp-guide-09-security">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 06 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>mcp</category>
      <category>typescript</category>
      <category>python</category>
    </item>
    <item>
      <title><![CDATA[8장: ref 개선, 메타데이터, 리소스 로딩 API]]></title>
      <link>https://archive.kreathlab.com/tech/react19-rsc-08-api-changes</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/react19-rsc-08-api-changes</guid>
      <description><![CDATA[React 19의 DX 개선사항을 다룹니다. ref를 일반 props로 전달하는 방법, 컴포넌트 내 메타데이터 태그, 리소스 프리로딩 API를 살펴봅니다.]]></description>
      <content:encoded><![CDATA[<p>React 19의 DX 개선사항을 다룹니다. ref를 일반 props로 전달하는 방법, 컴포넌트 내 메타데이터 태그, 리소스 프리로딩 API를 살펴봅니다.</p><p><strong>웹 개발</strong> / 16분 읽기 / 시리즈: react19-rsc</p><p><a href="https://archive.kreathlab.com/tech/react19-rsc-08-api-changes">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 06 Feb 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>react</category>
      <category>nextjs</category>
      <category>performance</category>
      <category>frontend</category>
      <category>typescript</category>
    </item>
    <item>
      <title><![CDATA[3장: GitHub Copilot 심층 분석]]></title>
      <link>https://archive.kreathlab.com/tech/ai-coding-assistant-03-copilot</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-coding-assistant-03-copilot</guid>
      <description><![CDATA[GitHub Copilot의 아키텍처, 코드 컴플리션부터 Agent Mode까지의 기능 진화, VS Code/JetBrains/CLI 통합, 가격 플랜을 심층 분석합니다.]]></description>
      <content:encoded><![CDATA[<p>GitHub Copilot의 아키텍처, 코드 컴플리션부터 Agent Mode까지의 기능 진화, VS Code/JetBrains/CLI 통합, 가격 플랜을 심층 분석합니다.</p><p><strong>개발 도구</strong> / 15분 읽기 / 시리즈: ai-coding-assistant</p><p><a href="https://archive.kreathlab.com/tech/ai-coding-assistant-03-copilot">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 05 Feb 2026 00:00:00 GMT</pubDate>
      <category>devtools</category>
      <category>ai-coding</category>
      <category>claude-code</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[10장: 실전 프로젝트 - 프로덕션 AI 서비스 파이프라인 구축]]></title>
      <link>https://archive.kreathlab.com/tech/ai-deployment-10-practical-project</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-deployment-10-practical-project</guid>
      <description><![CDATA[모델 서빙부터 Kubernetes 배포, 오토스케일링, CI/CD까지 전체 AI 서비스 배포 파이프라인을 처음부터 끝까지 구축하는 종합 실전 프로젝트입니다.]]></description>
      <content:encoded><![CDATA[<p>모델 서빙부터 Kubernetes 배포, 오토스케일링, CI/CD까지 전체 AI 서비스 배포 파이프라인을 처음부터 끝까지 구축하는 종합 실전 프로젝트입니다.</p><p><strong>AI / ML</strong> / 20분 읽기 / 시리즈: ai-deployment</p><p><a href="https://archive.kreathlab.com/tech/ai-deployment-10-practical-project">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 05 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>mlops</category>
      <category>kubernetes</category>
      <category>infrastructure</category>
      <category>performance</category>
    </item>
    <item>
      <title><![CDATA[2장: 시맨틱 검색 아키텍처]]></title>
      <link>https://archive.kreathlab.com/tech/ai-search-02-semantic-architecture</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-search-02-semantic-architecture</guid>
      <description><![CDATA[Bi-encoder 기반 시맨틱 검색의 작동 원리, 임베딩 모델 선택, 문서 청킹 전략, ANN 검색, 벡터 데이터베이스 연동을 Python 구현과 함께 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Bi-encoder 기반 시맨틱 검색의 작동 원리, 임베딩 모델 선택, 문서 청킹 전략, ANN 검색, 벡터 데이터베이스 연동을 Python 구현과 함께 다룹니다.</p><p><strong>AI / ML</strong> / 13분 읽기 / 시리즈: ai-search</p><p><a href="https://archive.kreathlab.com/tech/ai-search-02-semantic-architecture">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 05 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>search</category>
      <category>ai</category>
    </item>
    <item>
      <title><![CDATA[9장: Turbopack - 차세대 번들러의 시대]]></title>
      <link>https://archive.kreathlab.com/tech/nextjs-app-router-09-turbopack</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/nextjs-app-router-09-turbopack</guid>
      <description><![CDATA[Next.js의 차세대 번들러 Turbopack을 다룹니다. Rust 기반 아키텍처, 성능 벤치마크, FS 캐싱, Webpack 마이그레이션, 설정 방법을 살펴봅니다.]]></description>
      <content:encoded><![CDATA[<p>Next.js의 차세대 번들러 Turbopack을 다룹니다. Rust 기반 아키텍처, 성능 벤치마크, FS 캐싱, Webpack 마이그레이션, 설정 방법을 살펴봅니다.</p><p><strong>웹 개발</strong> / 20분 읽기 / 시리즈: nextjs-app-router</p><p><a href="https://archive.kreathlab.com/tech/nextjs-app-router-09-turbopack">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 05 Feb 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>nextjs</category>
      <category>react</category>
      <category>typescript</category>
      <category>frontend</category>
    </item>
    <item>
      <title><![CDATA[11장: typing 고급 기능 실전 활용]]></title>
      <link>https://archive.kreathlab.com/tech/python-trends-11-advanced-typing</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/python-trends-11-advanced-typing</guid>
      <description><![CDATA[Python 3.12~3.13의 typing 모듈 고급 기능을 다룹니다. TypedDict, Protocol, override, dataclass_transform, TypeGuard, TypeIs 등 실전 타입 시스템을 안내합니다.]]></description>
      <content:encoded><![CDATA[<p>Python 3.12~3.13의 typing 모듈 고급 기능을 다룹니다. TypedDict, Protocol, override, dataclass_transform, TypeGuard, TypeIs 등 실전 타입 시스템을 안내합니다.</p><p><strong>프로그래밍</strong> / 13분 읽기 / 시리즈: python-trends</p><p><a href="https://archive.kreathlab.com/tech/python-trends-11-advanced-typing">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 05 Feb 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>python</category>
      <category>typescript</category>
      <category>performance</category>
      <category>devtools</category>
      <category>concurrency</category>
    </item>
    <item>
      <title><![CDATA[8장: infer 키워드와 타입 추론 마스터]]></title>
      <link>https://archive.kreathlab.com/tech/typescript-deepdive-08-infer-patterns</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/typescript-deepdive-08-infer-patterns</guid>
      <description><![CDATA[TypeScript infer 키워드의 고급 활용 패턴, infer extends 구문, 공변/반변 위치에서의 추론, 그리고 실전 타입 추출 패턴을 심층 분석합니다.]]></description>
      <content:encoded><![CDATA[<p>TypeScript infer 키워드의 고급 활용 패턴, infer extends 구문, 공변/반변 위치에서의 추론, 그리고 실전 타입 추출 패턴을 심층 분석합니다.</p><p><strong>프로그래밍</strong> / 14분 읽기 / 시리즈: typescript-deepdive</p><p><a href="https://archive.kreathlab.com/tech/typescript-deepdive-08-infer-patterns">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 05 Feb 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>typescript</category>
      <category>performance</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[10장: 실전 프로젝트 - 리서치 에이전트 시스템 구축]]></title>
      <link>https://archive.kreathlab.com/tech/ai-agent-patterns-10-practical-project</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-agent-patterns-10-practical-project</guid>
      <description><![CDATA[이 시리즈에서 배운 모든 패턴을 결합하여 실제 사용 가능한 리서치 에이전트 시스템을 설계하고 구축합니다.]]></description>
      <content:encoded><![CDATA[<p>이 시리즈에서 배운 모든 패턴을 결합하여 실제 사용 가능한 리서치 에이전트 시스템을 설계하고 구축합니다.</p><p><strong>AI / ML</strong> / 21분 읽기 / 시리즈: ai-agent-patterns</p><p><a href="https://archive.kreathlab.com/tech/ai-agent-patterns-10-practical-project">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 04 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai-agent</category>
      <category>llm</category>
      <category>architecture</category>
      <category>orchestration</category>
    </item>
    <item>
      <title><![CDATA[9장: AI 통합 CI/CD 파이프라인 구축]]></title>
      <link>https://archive.kreathlab.com/tech/ai-dev-workflow-09-cicd-pipeline</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-dev-workflow-09-cicd-pipeline</guid>
      <description><![CDATA[코드 리뷰, 테스트 생성, 문서화, PR 분석을 하나의 CI/CD 파이프라인으로 통합하고, 품질 게이트와 비용 관리 전략을 수립합니다.]]></description>
      <content:encoded><![CDATA[<p>코드 리뷰, 테스트 생성, 문서화, PR 분석을 하나의 CI/CD 파이프라인으로 통합하고, 품질 게이트와 비용 관리 전략을 수립합니다.</p><p><strong>AI / ML</strong> / 18분 읽기 / 시리즈: ai-dev-workflow</p><p><a href="https://archive.kreathlab.com/tech/ai-dev-workflow-09-cicd-pipeline">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 04 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>devtools</category>
      <category>automation</category>
      <category>code-quality</category>
      <category>devops</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[2장: LangChain 아키텍처와 LCEL 심층 분석]]></title>
      <link>https://archive.kreathlab.com/tech/ai-orchestration-02-langchain</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-orchestration-02-langchain</guid>
      <description><![CDATA[LangChain 1.0의 아키텍처, LCEL 파이프 문법, 미들웨어, 콘텐츠 블록, OpenTelemetry 통합을 실전 예제와 함께 분석합니다.]]></description>
      <content:encoded><![CDATA[<p>LangChain 1.0의 아키텍처, LCEL 파이프 문법, 미들웨어, 콘텐츠 블록, OpenTelemetry 통합을 실전 예제와 함께 분석합니다.</p><p><strong>AI / ML</strong> / 14분 읽기 / 시리즈: ai-orchestration</p><p><a href="https://archive.kreathlab.com/tech/ai-orchestration-02-langchain">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 04 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>orchestration</category>
      <category>ai-framework</category>
      <category>ai</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[1장: API 설계의 진화와 AI 서비스의 도전]]></title>
      <link>https://archive.kreathlab.com/tech/api-design-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/api-design-01-introduction</guid>
      <description><![CDATA[SOAP에서 REST, GraphQL, gRPC까지 API 패러다임의 진화를 살펴보고, AI 서비스가 직면한 고유 과제와 2026년 하이브리드 아키텍처 트렌드를 분석합니다.]]></description>
      <content:encoded><![CDATA[<p>SOAP에서 REST, GraphQL, gRPC까지 API 패러다임의 진화를 살펴보고, AI 서비스가 직면한 고유 과제와 2026년 하이브리드 아키텍처 트렌드를 분석합니다.</p><p><strong>아키텍처</strong> / 17분 읽기 / 시리즈: api-design</p><p><a href="https://archive.kreathlab.com/tech/api-design-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 04 Feb 2026 00:00:00 GMT</pubDate>
      <category>architecture</category>
      <category>api-design</category>
      <category>graphql</category>
      <category>architecture</category>
    </item>
    <item>
      <title><![CDATA[8장: 기존 시스템과 MCP 연동하기]]></title>
      <link>https://archive.kreathlab.com/tech/mcp-guide-08-system-integration</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/mcp-guide-08-system-integration</guid>
      <description><![CDATA[데이터베이스, REST API, 레거시 시스템을 MCP 서버로 래핑하여 AI 모델이 접근할 수 있도록 만드는 실전 패턴을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>데이터베이스, REST API, 레거시 시스템을 MCP 서버로 래핑하여 AI 모델이 접근할 수 있도록 만드는 실전 패턴을 다룹니다.</p><p><strong>AI / ML</strong> / 18분 읽기 / 시리즈: mcp-guide</p><p><a href="https://archive.kreathlab.com/tech/mcp-guide-08-system-integration">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 04 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>mcp</category>
      <category>typescript</category>
      <category>python</category>
    </item>
    <item>
      <title><![CDATA[7장: React Compiler - 자동 최적화의 시대]]></title>
      <link>https://archive.kreathlab.com/tech/react19-rsc-07-compiler</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/react19-rsc-07-compiler</guid>
      <description><![CDATA[React Compiler의 동작 원리, HIR 기반 분석, 자동 메모이제이션, 설치와 설정, ESLint 통합, 실전 적용 전략을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>React Compiler의 동작 원리, HIR 기반 분석, 자동 메모이제이션, 설치와 설정, ESLint 통합, 실전 적용 전략을 다룹니다.</p><p><strong>웹 개발</strong> / 15분 읽기 / 시리즈: react19-rsc</p><p><a href="https://archive.kreathlab.com/tech/react19-rsc-07-compiler">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 04 Feb 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>react</category>
      <category>nextjs</category>
      <category>performance</category>
      <category>frontend</category>
      <category>typescript</category>
    </item>
    <item>
      <title><![CDATA[2장: 내부 동작 원리 — LLM 기반 코드 생성]]></title>
      <link>https://archive.kreathlab.com/tech/ai-coding-assistant-02-internals</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-coding-assistant-02-internals</guid>
      <description><![CDATA[Fill-in-the-Middle, 코드 LLM 학습 방법, 토큰화, AST 인식 등 AI 코딩 어시스턴트의 핵심 기술 원리를 깊이 있게 분석합니다.]]></description>
      <content:encoded><![CDATA[<p>Fill-in-the-Middle, 코드 LLM 학습 방법, 토큰화, AST 인식 등 AI 코딩 어시스턴트의 핵심 기술 원리를 깊이 있게 분석합니다.</p><p><strong>개발 도구</strong> / 19분 읽기 / 시리즈: ai-coding-assistant</p><p><a href="https://archive.kreathlab.com/tech/ai-coding-assistant-02-internals">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 03 Feb 2026 00:00:00 GMT</pubDate>
      <category>devtools</category>
      <category>ai-coding</category>
      <category>claude-code</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[9장: CI/CD 파이프라인 - GitHub Actions로 모델 배포 자동화]]></title>
      <link>https://archive.kreathlab.com/tech/ai-deployment-09-cicd</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-deployment-09-cicd</guid>
      <description><![CDATA[GitHub Actions를 활용하여 AI 서비스의 빌드, 테스트, 배포를 자동화하는 CI/CD 파이프라인을 구축하고, 모델 평가를 파이프라인에 통합합니다.]]></description>
      <content:encoded><![CDATA[<p>GitHub Actions를 활용하여 AI 서비스의 빌드, 테스트, 배포를 자동화하는 CI/CD 파이프라인을 구축하고, 모델 평가를 파이프라인에 통합합니다.</p><p><strong>AI / ML</strong> / 14분 읽기 / 시리즈: ai-deployment</p><p><a href="https://archive.kreathlab.com/tech/ai-deployment-09-cicd">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 03 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>mlops</category>
      <category>kubernetes</category>
      <category>infrastructure</category>
      <category>performance</category>
    </item>
    <item>
      <title><![CDATA[1장: AI 기반 검색의 진화와 핵심 개념]]></title>
      <link>https://archive.kreathlab.com/tech/ai-search-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-search-01-introduction</guid>
      <description><![CDATA[키워드 검색에서 시맨틱 검색, 하이브리드 검색으로 이어지는 검색 기술의 진화 과정과 AI 검색 시스템의 핵심 구성요소를 살펴봅니다.]]></description>
      <content:encoded><![CDATA[<p>키워드 검색에서 시맨틱 검색, 하이브리드 검색으로 이어지는 검색 기술의 진화 과정과 AI 검색 시스템의 핵심 구성요소를 살펴봅니다.</p><p><strong>AI / ML</strong> / 15분 읽기 / 시리즈: ai-search</p><p><a href="https://archive.kreathlab.com/tech/ai-search-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 03 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>search</category>
      <category>ai</category>
    </item>
    <item>
      <title><![CDATA[8장: Server Actions와 폼 처리 고급 패턴]]></title>
      <link>https://archive.kreathlab.com/tech/nextjs-app-router-08-server-actions</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/nextjs-app-router-08-server-actions</guid>
      <description><![CDATA[Next.js의 Server Actions와 폼 처리를 다룹니다. next/form, useActionState, 보안, revalidation 통합, after() API, 고급 뮤테이션 패턴을 살펴봅니다.]]></description>
      <content:encoded><![CDATA[<p>Next.js의 Server Actions와 폼 처리를 다룹니다. next/form, useActionState, 보안, revalidation 통합, after() API, 고급 뮤테이션 패턴을 살펴봅니다.</p><p><strong>웹 개발</strong> / 21분 읽기 / 시리즈: nextjs-app-router</p><p><a href="https://archive.kreathlab.com/tech/nextjs-app-router-08-server-actions">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 03 Feb 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>nextjs</category>
      <category>react</category>
      <category>typescript</category>
      <category>frontend</category>
    </item>
    <item>
      <title><![CDATA[10장: Ruff와 ty - 차세대 린터, 포매터, 타입 체커]]></title>
      <link>https://archive.kreathlab.com/tech/python-trends-10-ruff-ty</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/python-trends-10-ruff-ty</guid>
      <description><![CDATA[Astral의 Ruff(린터/포매터)와 ty(타입 체커)를 다룹니다. 기존 도구 대체, 설정 방법, 규칙 커스터마이징, IDE 통합, 프로젝트 도입 전략을 안내합니다.]]></description>
      <content:encoded><![CDATA[<p>Astral의 Ruff(린터/포매터)와 ty(타입 체커)를 다룹니다. 기존 도구 대체, 설정 방법, 규칙 커스터마이징, IDE 통합, 프로젝트 도입 전략을 안내합니다.</p><p><strong>프로그래밍</strong> / 14분 읽기 / 시리즈: python-trends</p><p><a href="https://archive.kreathlab.com/tech/python-trends-10-ruff-ty">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 03 Feb 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>python</category>
      <category>typescript</category>
      <category>performance</category>
      <category>devtools</category>
      <category>concurrency</category>
    </item>
    <item>
      <title><![CDATA[7장: 템플릿 리터럴 타입으로 문자열 다루기]]></title>
      <link>https://archive.kreathlab.com/tech/typescript-deepdive-07-template-literal-types</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/typescript-deepdive-07-template-literal-types</guid>
      <description><![CDATA[TypeScript 템플릿 리터럴 타입의 원리, 내장 문자열 조작 유틸리티, 패턴 매칭, 그리고 매핑 타입과의 결합을 통한 실전 패턴을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>TypeScript 템플릿 리터럴 타입의 원리, 내장 문자열 조작 유틸리티, 패턴 매칭, 그리고 매핑 타입과의 결합을 통한 실전 패턴을 다룹니다.</p><p><strong>프로그래밍</strong> / 12분 읽기 / 시리즈: typescript-deepdive</p><p><a href="https://archive.kreathlab.com/tech/typescript-deepdive-07-template-literal-types">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 03 Feb 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>typescript</category>
      <category>performance</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[9장: 에이전트 프레임워크 비교 - LangGraph, CrewAI, OpenAI Agents SDK]]></title>
      <link>https://archive.kreathlab.com/tech/ai-agent-patterns-09-frameworks</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-agent-patterns-09-frameworks</guid>
      <description><![CDATA[주요 AI 에이전트 프레임워크의 아키텍처, 장단점, 사용 사례를 비교하고 프로젝트에 적합한 프레임워크를 선택하는 기준을 제시합니다.]]></description>
      <content:encoded><![CDATA[<p>주요 AI 에이전트 프레임워크의 아키텍처, 장단점, 사용 사례를 비교하고 프로젝트에 적합한 프레임워크를 선택하는 기준을 제시합니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: ai-agent-patterns</p><p><a href="https://archive.kreathlab.com/tech/ai-agent-patterns-09-frameworks">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 02 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai-agent</category>
      <category>llm</category>
      <category>architecture</category>
      <category>orchestration</category>
    </item>
    <item>
      <title><![CDATA[8장: Claude Code를 활용한 개발 자동화]]></title>
      <link>https://archive.kreathlab.com/tech/ai-dev-workflow-08-claude-code</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-dev-workflow-08-claude-code</guid>
      <description><![CDATA[Claude Code의 에이전트 기반 워크플로우를 활용하여 코드 생성, 리팩터링, 디버깅을 자동화하고, CI/CD에 통합하는 방법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Claude Code의 에이전트 기반 워크플로우를 활용하여 코드 생성, 리팩터링, 디버깅을 자동화하고, CI/CD에 통합하는 방법을 다룹니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: ai-dev-workflow</p><p><a href="https://archive.kreathlab.com/tech/ai-dev-workflow-08-claude-code">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 02 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>devtools</category>
      <category>automation</category>
      <category>code-quality</category>
      <category>devops</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[1장: AI 오케스트레이션의 필요성과 프레임워크 생태계]]></title>
      <link>https://archive.kreathlab.com/tech/ai-orchestration-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-orchestration-01-introduction</guid>
      <description><![CDATA[LLM 애플리케이션이 복잡해지는 이유를 분석하고, 오케스트레이션의 정의와 역할, 2026년 주요 프레임워크 생태계를 조망합니다.]]></description>
      <content:encoded><![CDATA[<p>LLM 애플리케이션이 복잡해지는 이유를 분석하고, 오케스트레이션의 정의와 역할, 2026년 주요 프레임워크 생태계를 조망합니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: ai-orchestration</p><p><a href="https://archive.kreathlab.com/tech/ai-orchestration-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 02 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>orchestration</category>
      <category>ai-framework</category>
      <category>ai</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[10장: 실전 프로젝트 - 종합 평가 모니터링 시스템 구축]]></title>
      <link>https://archive.kreathlab.com/tech/llm-evaluation-10-practical-project</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/llm-evaluation-10-practical-project</guid>
      <description><![CDATA[지금까지 다룬 평가 메트릭, LLM-as-Judge, 모니터링, CI/CD를 통합하여 프로덕션 수준의 종합 평가 시스템을 구축합니다.]]></description>
      <content:encoded><![CDATA[<p>지금까지 다룬 평가 메트릭, LLM-as-Judge, 모니터링, CI/CD를 통합하여 프로덕션 수준의 종합 평가 시스템을 구축합니다.</p><p><strong>AI / ML</strong> / 22분 읽기 / 시리즈: llm-evaluation</p><p><a href="https://archive.kreathlab.com/tech/llm-evaluation-10-practical-project">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 02 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>evaluation</category>
      <category>monitoring</category>
      <category>observability</category>
      <category>testing</category>
    </item>
    <item>
      <title><![CDATA[7장: MCP 클라이언트 구현하기]]></title>
      <link>https://archive.kreathlab.com/tech/mcp-guide-07-client-implementation</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/mcp-guide-07-client-implementation</guid>
      <description><![CDATA[MCP 클라이언트를 직접 구현하여 서버에 연결하고, LLM과 통합하여 도구 호출 파이프라인을 완성하는 방법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>MCP 클라이언트를 직접 구현하여 서버에 연결하고, LLM과 통합하여 도구 호출 파이프라인을 완성하는 방법을 다룹니다.</p><p><strong>AI / ML</strong> / 15분 읽기 / 시리즈: mcp-guide</p><p><a href="https://archive.kreathlab.com/tech/mcp-guide-07-client-implementation">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 02 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>mcp</category>
      <category>typescript</category>
      <category>python</category>
    </item>
    <item>
      <title><![CDATA[6장: Suspense 고급 패턴과 스트리밍 SSR]]></title>
      <link>https://archive.kreathlab.com/tech/react19-rsc-06-suspense</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/react19-rsc-06-suspense</guid>
      <description><![CDATA[React 19에서 강화된 Suspense의 고급 패턴, 스트리밍 SSR, 중첩 Suspense 전략, 배칭 동작, Partial Pre-rendering을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>React 19에서 강화된 Suspense의 고급 패턴, 스트리밍 SSR, 중첩 Suspense 전략, 배칭 동작, Partial Pre-rendering을 다룹니다.</p><p><strong>웹 개발</strong> / 16분 읽기 / 시리즈: react19-rsc</p><p><a href="https://archive.kreathlab.com/tech/react19-rsc-06-suspense">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 02 Feb 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>react</category>
      <category>nextjs</category>
      <category>performance</category>
      <category>frontend</category>
      <category>typescript</category>
    </item>
    <item>
      <title><![CDATA[1장: AI 코딩 어시스턴트의 등장과 현재]]></title>
      <link>https://archive.kreathlab.com/tech/ai-coding-assistant-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-coding-assistant-01-introduction</guid>
      <description><![CDATA[2026년 현재 84%의 개발자가 사용하는 AI 코딩 어시스턴트의 현황과 3가지 아키텍처 철학, 주요 도구 지도, 그리고 생산성 논쟁을 살펴봅니다.]]></description>
      <content:encoded><![CDATA[<p>2026년 현재 84%의 개발자가 사용하는 AI 코딩 어시스턴트의 현황과 3가지 아키텍처 철학, 주요 도구 지도, 그리고 생산성 논쟁을 살펴봅니다.</p><p><strong>개발 도구</strong> / 15분 읽기 / 시리즈: ai-coding-assistant</p><p><a href="https://archive.kreathlab.com/tech/ai-coding-assistant-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 01 Feb 2026 00:00:00 GMT</pubDate>
      <category>devtools</category>
      <category>ai-coding</category>
      <category>claude-code</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[8장: 비용 최적화 - 스팟 인스턴스, 모델 공유, 리소스 관리]]></title>
      <link>https://archive.kreathlab.com/tech/ai-deployment-08-cost-optimization</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-deployment-08-cost-optimization</guid>
      <description><![CDATA[GPU 기반 AI 서비스의 운영 비용을 체계적으로 절감하는 전략을 다루며, 스팟 인스턴스 활용, 모델 공유 아키텍처, 리소스 관리 기법을 소개합니다.]]></description>
      <content:encoded><![CDATA[<p>GPU 기반 AI 서비스의 운영 비용을 체계적으로 절감하는 전략을 다루며, 스팟 인스턴스 활용, 모델 공유 아키텍처, 리소스 관리 기법을 소개합니다.</p><p><strong>AI / ML</strong> / 18분 읽기 / 시리즈: ai-deployment</p><p><a href="https://archive.kreathlab.com/tech/ai-deployment-08-cost-optimization">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 01 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>mlops</category>
      <category>kubernetes</category>
      <category>infrastructure</category>
      <category>performance</category>
    </item>
    <item>
      <title><![CDATA[10장: 실전 프로젝트 - 도메인 특화 코드 리뷰 모델 파인튜닝]]></title>
      <link>https://archive.kreathlab.com/tech/fine-tuning-10-practical-project</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/fine-tuning-10-practical-project</guid>
      <description><![CDATA[코드 리뷰 특화 모델을 데이터 수집부터 프로덕션 배포까지 전 과정을 실습하며, 시리즈에서 배운 모든 기법을 통합 적용합니다.]]></description>
      <content:encoded><![CDATA[<p>코드 리뷰 특화 모델을 데이터 수집부터 프로덕션 배포까지 전 과정을 실습하며, 시리즈에서 배운 모든 기법을 통합 적용합니다.</p><p><strong>AI / ML</strong> / 20분 읽기 / 시리즈: fine-tuning</p><p><a href="https://archive.kreathlab.com/tech/fine-tuning-10-practical-project">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 01 Feb 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>training</category>
      <category>mlops</category>
      <category>data-engineering</category>
    </item>
    <item>
      <title><![CDATA[7장: 스트리밍 SSR과 로딩 UI 전략]]></title>
      <link>https://archive.kreathlab.com/tech/nextjs-app-router-07-streaming</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/nextjs-app-router-07-streaming</guid>
      <description><![CDATA[Next.js의 스트리밍 SSR 동작 원리를 살펴봅니다. loading.tsx, Suspense 경계, 스켈레톤 설계, 프로그레시브 렌더링 전략과 성능 지표 영향을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Next.js의 스트리밍 SSR 동작 원리를 살펴봅니다. loading.tsx, Suspense 경계, 스켈레톤 설계, 프로그레시브 렌더링 전략과 성능 지표 영향을 다룹니다.</p><p><strong>웹 개발</strong> / 17분 읽기 / 시리즈: nextjs-app-router</p><p><a href="https://archive.kreathlab.com/tech/nextjs-app-router-07-streaming">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 01 Feb 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>nextjs</category>
      <category>react</category>
      <category>typescript</category>
      <category>frontend</category>
    </item>
    <item>
      <title><![CDATA[9장: uv - 차세대 Python 패키지 매니저]]></title>
      <link>https://archive.kreathlab.com/tech/python-trends-09-uv</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/python-trends-09-uv</guid>
      <description><![CDATA[Rust로 작성된 차세대 Python 패키지 매니저 uv를 다룹니다. 설치, 프로젝트 관리, 가상 환경, Python 버전 관리, 스크립트 실행까지 실전 워크플로우를 안내합니다.]]></description>
      <content:encoded><![CDATA[<p>Rust로 작성된 차세대 Python 패키지 매니저 uv를 다룹니다. 설치, 프로젝트 관리, 가상 환경, Python 버전 관리, 스크립트 실행까지 실전 워크플로우를 안내합니다.</p><p><strong>프로그래밍</strong> / 14분 읽기 / 시리즈: python-trends</p><p><a href="https://archive.kreathlab.com/tech/python-trends-09-uv">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 01 Feb 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>python</category>
      <category>typescript</category>
      <category>performance</category>
      <category>devtools</category>
      <category>concurrency</category>
    </item>
    <item>
      <title><![CDATA[6장: 매핑 타입과 키 재매핑 고급 패턴]]></title>
      <link>https://archive.kreathlab.com/tech/typescript-deepdive-06-mapped-types</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/typescript-deepdive-06-mapped-types</guid>
      <description><![CDATA[TypeScript 매핑 타입의 원리, 수정자 조작, 키 재매핑(as 절), 그리고 조건부 타입과의 결합 패턴을 실전 예제와 함께 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>TypeScript 매핑 타입의 원리, 수정자 조작, 키 재매핑(as 절), 그리고 조건부 타입과의 결합 패턴을 실전 예제와 함께 다룹니다.</p><p><strong>프로그래밍</strong> / 12분 읽기 / 시리즈: typescript-deepdive</p><p><a href="https://archive.kreathlab.com/tech/typescript-deepdive-06-mapped-types">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 01 Feb 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>typescript</category>
      <category>performance</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[8장: 가드레일과 안전성 - 에이전트를 신뢰할 수 있게 만들기]]></title>
      <link>https://archive.kreathlab.com/tech/ai-agent-patterns-08-guardrails</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-agent-patterns-08-guardrails</guid>
      <description><![CDATA[AI 에이전트의 행동 제어, 입출력 검증, 오류 처리, 비용 관리 등 프로덕션 환경에서의 안전성 확보 전략을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>AI 에이전트의 행동 제어, 입출력 검증, 오류 처리, 비용 관리 등 프로덕션 환경에서의 안전성 확보 전략을 다룹니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: ai-agent-patterns</p><p><a href="https://archive.kreathlab.com/tech/ai-agent-patterns-08-guardrails">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 31 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai-agent</category>
      <category>llm</category>
      <category>architecture</category>
      <category>orchestration</category>
    </item>
    <item>
      <title><![CDATA[7장: GitHub Copilot 심층 활용 전략]]></title>
      <link>https://archive.kreathlab.com/tech/ai-dev-workflow-07-copilot</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-dev-workflow-07-copilot</guid>
      <description><![CDATA[GitHub Copilot의 인라인 자동 완성, Copilot Chat, Agent Mode를 실전에서 효과적으로 활용하는 전략과 팀 단위 도입 방법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>GitHub Copilot의 인라인 자동 완성, Copilot Chat, Agent Mode를 실전에서 효과적으로 활용하는 전략과 팀 단위 도입 방법을 다룹니다.</p><p><strong>AI / ML</strong> / 19분 읽기 / 시리즈: ai-dev-workflow</p><p><a href="https://archive.kreathlab.com/tech/ai-dev-workflow-07-copilot">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 31 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>devtools</category>
      <category>automation</category>
      <category>code-quality</category>
      <category>devops</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[9장: CI/CD에 평가 파이프라인 통합]]></title>
      <link>https://archive.kreathlab.com/tech/llm-evaluation-09-cicd-integration</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/llm-evaluation-09-cicd-integration</guid>
      <description><![CDATA[LLM 평가를 CI/CD 파이프라인에 통합하여, 프롬프트 변경과 모델 교체 시 자동으로 품질을 검증하는 체계를 구축합니다.]]></description>
      <content:encoded><![CDATA[<p>LLM 평가를 CI/CD 파이프라인에 통합하여, 프롬프트 변경과 모델 교체 시 자동으로 품질을 검증하는 체계를 구축합니다.</p><p><strong>AI / ML</strong> / 15분 읽기 / 시리즈: llm-evaluation</p><p><a href="https://archive.kreathlab.com/tech/llm-evaluation-09-cicd-integration">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 31 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>evaluation</category>
      <category>monitoring</category>
      <category>observability</category>
      <category>testing</category>
    </item>
    <item>
      <title><![CDATA[6장: Python FastMCP로 서버 구축하기]]></title>
      <link>https://archive.kreathlab.com/tech/mcp-guide-06-python-fastmcp</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/mcp-guide-06-python-fastmcp</guid>
      <description><![CDATA[Python의 FastMCP 프레임워크를 사용하여 데코레이터 기반의 간결하고 직관적인 MCP 서버를 구축하는 방법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Python의 FastMCP 프레임워크를 사용하여 데코레이터 기반의 간결하고 직관적인 MCP 서버를 구축하는 방법을 다룹니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: mcp-guide</p><p><a href="https://archive.kreathlab.com/tech/mcp-guide-06-python-fastmcp">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 31 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>mcp</category>
      <category>typescript</category>
      <category>python</category>
    </item>
    <item>
      <title><![CDATA[10장: 프로덕션 RAG 파이프라인 구축]]></title>
      <link>https://archive.kreathlab.com/tech/rag-system-10-production</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/rag-system-10-production</guid>
      <description><![CDATA[모니터링, 캐싱, 보안, 확장성, 배포 전략까지 프로덕션 수준의 RAG 시스템을 설계하고 운영하는 실전 가이드입니다.]]></description>
      <content:encoded><![CDATA[<p>모니터링, 캐싱, 보안, 확장성, 배포 전략까지 프로덕션 수준의 RAG 시스템을 설계하고 운영하는 실전 가이드입니다.</p><p><strong>AI / ML</strong> / 18분 읽기 / 시리즈: rag-system</p><p><a href="https://archive.kreathlab.com/tech/rag-system-10-production">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 31 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>rag</category>
      <category>vector-database</category>
      <category>embedding</category>
      <category>retrieval</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[5장: 새로운 훅 - useActionState, useFormStatus, useOptimistic]]></title>
      <link>https://archive.kreathlab.com/tech/react19-rsc-05-new-hooks</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/react19-rsc-05-new-hooks</guid>
      <description><![CDATA[React 19의 새로운 훅 3종을 심층 분석합니다. 폼 상태 관리, 제출 상태 추적, 낙관적 UI 업데이트의 실전 패턴을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>React 19의 새로운 훅 3종을 심층 분석합니다. 폼 상태 관리, 제출 상태 추적, 낙관적 UI 업데이트의 실전 패턴을 다룹니다.</p><p><strong>웹 개발</strong> / 16분 읽기 / 시리즈: react19-rsc</p><p><a href="https://archive.kreathlab.com/tech/react19-rsc-05-new-hooks">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 31 Jan 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>react</category>
      <category>nextjs</category>
      <category>performance</category>
      <category>frontend</category>
      <category>typescript</category>
    </item>
    <item>
      <title><![CDATA[7장: 오토스케일링 - 트래픽 기반 GPU 워크로드 확장]]></title>
      <link>https://archive.kreathlab.com/tech/ai-deployment-07-autoscaling</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-deployment-07-autoscaling</guid>
      <description><![CDATA[Kubernetes에서 GPU 기반 AI 서비스의 자동 확장 전략을 구현하며, HPA 커스텀 메트릭과 Cluster Autoscaler를 활용한 효율적인 스케일링 방법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Kubernetes에서 GPU 기반 AI 서비스의 자동 확장 전략을 구현하며, HPA 커스텀 메트릭과 Cluster Autoscaler를 활용한 효율적인 스케일링 방법을 다룹니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: ai-deployment</p><p><a href="https://archive.kreathlab.com/tech/ai-deployment-07-autoscaling">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 30 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>mlops</category>
      <category>kubernetes</category>
      <category>infrastructure</category>
      <category>performance</category>
    </item>
    <item>
      <title><![CDATA[9장: 학습-평가-배포 자동화 사이클]]></title>
      <link>https://archive.kreathlab.com/tech/fine-tuning-09-automation</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/fine-tuning-09-automation</guid>
      <description><![CDATA[파인튜닝의 학습, 평가, 배포 전체 과정을 CI/CD 파이프라인으로 자동화하고, 데이터 변경이나 코드 변경 시 자동으로 모델이 업데이트되는 체계를 구축합니다.]]></description>
      <content:encoded><![CDATA[<p>파인튜닝의 학습, 평가, 배포 전체 과정을 CI/CD 파이프라인으로 자동화하고, 데이터 변경이나 코드 변경 시 자동으로 모델이 업데이트되는 체계를 구축합니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: fine-tuning</p><p><a href="https://archive.kreathlab.com/tech/fine-tuning-09-automation">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 30 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>training</category>
      <category>mlops</category>
      <category>data-engineering</category>
    </item>
    <item>
      <title><![CDATA[6장: Cache Components와 'use cache' 디렉티브]]></title>
      <link>https://archive.kreathlab.com/tech/nextjs-app-router-06-cache-components</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/nextjs-app-router-06-cache-components</guid>
      <description><![CDATA[Next.js 16의 Cache Components 시스템을 다룹니다. 'use cache' 디렉티브, cacheLife 프로필, cacheTag 무효화, 세 가지 캐시 변형을 살펴봅니다.]]></description>
      <content:encoded><![CDATA[<p>Next.js 16의 Cache Components 시스템을 다룹니다. &apos;use cache&apos; 디렉티브, cacheLife 프로필, cacheTag 무효화, 세 가지 캐시 변형을 살펴봅니다.</p><p><strong>웹 개발</strong> / 17분 읽기 / 시리즈: nextjs-app-router</p><p><a href="https://archive.kreathlab.com/tech/nextjs-app-router-06-cache-components">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 30 Jan 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>nextjs</category>
      <category>react</category>
      <category>typescript</category>
      <category>frontend</category>
    </item>
    <item>
      <title><![CDATA[10장: 프로덕션 프롬프트 관리 - 버전 관리와 CI/CD]]></title>
      <link>https://archive.kreathlab.com/tech/prompt-engineering-10-production</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/prompt-engineering-10-production</guid>
      <description><![CDATA[프롬프트의 버전 관리, CI/CD 파이프라인 통합, 환경별 배포 전략, 그리고 운영 모니터링까지 프로덕션급 프롬프트 관리 체계를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>프롬프트의 버전 관리, CI/CD 파이프라인 통합, 환경별 배포 전략, 그리고 운영 모니터링까지 프로덕션급 프롬프트 관리 체계를 다룹니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: prompt-engineering</p><p><a href="https://archive.kreathlab.com/tech/prompt-engineering-10-production">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 30 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>prompt-engineering</category>
      <category>structured-output</category>
      <category>training</category>
    </item>
    <item>
      <title><![CDATA[8장: 실험적 JIT 컴파일러의 원리와 가능성]]></title>
      <link>https://archive.kreathlab.com/tech/python-trends-08-jit-compiler</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/python-trends-08-jit-compiler</guid>
      <description><![CDATA[Python 3.13에 도입된 실험적 JIT 컴파일러를 분석합니다. copy-and-patch 기법의 원리, Tier 2 최적화 파이프라인, 빌드와 활성화 방법, 성능 벤치마크를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Python 3.13에 도입된 실험적 JIT 컴파일러를 분석합니다. copy-and-patch 기법의 원리, Tier 2 최적화 파이프라인, 빌드와 활성화 방법, 성능 벤치마크를 다룹니다.</p><p><strong>프로그래밍</strong> / 17분 읽기 / 시리즈: python-trends</p><p><a href="https://archive.kreathlab.com/tech/python-trends-08-jit-compiler">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 30 Jan 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>python</category>
      <category>typescript</category>
      <category>performance</category>
      <category>devtools</category>
      <category>concurrency</category>
    </item>
    <item>
      <title><![CDATA[5장: 조건부 타입 심층 분석]]></title>
      <link>https://archive.kreathlab.com/tech/typescript-deepdive-05-conditional-types</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/typescript-deepdive-05-conditional-types</guid>
      <description><![CDATA[TypeScript 조건부 타입의 원리, 분배적 조건부 타입, infer 키워드와의 조합, 그리고 실전 활용 패턴을 깊이 있게 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>TypeScript 조건부 타입의 원리, 분배적 조건부 타입, infer 키워드와의 조합, 그리고 실전 활용 패턴을 깊이 있게 다룹니다.</p><p><strong>프로그래밍</strong> / 13분 읽기 / 시리즈: typescript-deepdive</p><p><a href="https://archive.kreathlab.com/tech/typescript-deepdive-05-conditional-types">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 30 Jan 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>typescript</category>
      <category>performance</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[7장: 메모리 시스템 - 에이전트의 기억과 학습]]></title>
      <link>https://archive.kreathlab.com/tech/ai-agent-patterns-07-memory</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-agent-patterns-07-memory</guid>
      <description><![CDATA[AI 에이전트의 단기, 장기 메모리 아키텍처를 이해하고, RAG 통합과 대화 히스토리 관리 전략을 코드로 구현합니다.]]></description>
      <content:encoded><![CDATA[<p>AI 에이전트의 단기, 장기 메모리 아키텍처를 이해하고, RAG 통합과 대화 히스토리 관리 전략을 코드로 구현합니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: ai-agent-patterns</p><p><a href="https://archive.kreathlab.com/tech/ai-agent-patterns-07-memory">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 29 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai-agent</category>
      <category>llm</category>
      <category>architecture</category>
      <category>orchestration</category>
    </item>
    <item>
      <title><![CDATA[6장: PR 분석과 변경 영향도 예측]]></title>
      <link>https://archive.kreathlab.com/tech/ai-dev-workflow-06-pr-analysis</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-dev-workflow-06-pr-analysis</guid>
      <description><![CDATA[PR의 변경 범위와 위험도를 AI로 분석하고, 리뷰어에게 구조화된 인사이트를 제공하는 시스템을 구축합니다.]]></description>
      <content:encoded><![CDATA[<p>PR의 변경 범위와 위험도를 AI로 분석하고, 리뷰어에게 구조화된 인사이트를 제공하는 시스템을 구축합니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: ai-dev-workflow</p><p><a href="https://archive.kreathlab.com/tech/ai-dev-workflow-06-pr-analysis">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 29 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>devtools</category>
      <category>automation</category>
      <category>code-quality</category>
      <category>devops</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[8장: 드리프트 감지와 품질 모니터링]]></title>
      <link>https://archive.kreathlab.com/tech/llm-evaluation-08-drift-detection</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/llm-evaluation-08-drift-detection</guid>
      <description><![CDATA[LLM 애플리케이션의 입력 분포 변화, 모델 성능 저하, 데이터 드리프트를 감지하고 대응하는 방법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>LLM 애플리케이션의 입력 분포 변화, 모델 성능 저하, 데이터 드리프트를 감지하고 대응하는 방법을 다룹니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: llm-evaluation</p><p><a href="https://archive.kreathlab.com/tech/llm-evaluation-08-drift-detection">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 29 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>evaluation</category>
      <category>monitoring</category>
      <category>observability</category>
      <category>testing</category>
    </item>
    <item>
      <title><![CDATA[5장: TypeScript로 MCP 서버 구축하기]]></title>
      <link>https://archive.kreathlab.com/tech/mcp-guide-05-typescript-server</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/mcp-guide-05-typescript-server</guid>
      <description><![CDATA[TypeScript SDK를 사용하여 프로젝트 설정부터 도구, 리소스, 프롬프트 구현, 테스트까지 MCP 서버를 구축하는 전 과정을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>TypeScript SDK를 사용하여 프로젝트 설정부터 도구, 리소스, 프롬프트 구현, 테스트까지 MCP 서버를 구축하는 전 과정을 다룹니다.</p><p><strong>AI / ML</strong> / 18분 읽기 / 시리즈: mcp-guide</p><p><a href="https://archive.kreathlab.com/tech/mcp-guide-05-typescript-server">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 29 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>mcp</category>
      <category>typescript</category>
      <category>python</category>
    </item>
    <item>
      <title><![CDATA[9장: 고급 RAG 패턴 - Agentic RAG와 Self-Correcting RAG]]></title>
      <link>https://archive.kreathlab.com/tech/rag-system-09-advanced-patterns</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/rag-system-09-advanced-patterns</guid>
      <description><![CDATA[에이전트가 검색 전략을 스스로 판단하고 실패를 자동 수정하는 Agentic RAG, CRAG, Self-RAG 등 고급 RAG 패턴을 심층 분석합니다.]]></description>
      <content:encoded><![CDATA[<p>에이전트가 검색 전략을 스스로 판단하고 실패를 자동 수정하는 Agentic RAG, CRAG, Self-RAG 등 고급 RAG 패턴을 심층 분석합니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: rag-system</p><p><a href="https://archive.kreathlab.com/tech/rag-system-09-advanced-patterns">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 29 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>rag</category>
      <category>vector-database</category>
      <category>embedding</category>
      <category>retrieval</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[4장: use() API와 새로운 데이터 패턴]]></title>
      <link>https://archive.kreathlab.com/tech/react19-rsc-04-use-api</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/react19-rsc-04-use-api</guid>
      <description><![CDATA[React 19의 use() API로 Promise와 Context를 조건부로 소비하는 방법, 서버-클라이언트 스트리밍 패턴, 기존 훅과의 차이를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>React 19의 use() API로 Promise와 Context를 조건부로 소비하는 방법, 서버-클라이언트 스트리밍 패턴, 기존 훅과의 차이를 다룹니다.</p><p><strong>웹 개발</strong> / 15분 읽기 / 시리즈: react19-rsc</p><p><a href="https://archive.kreathlab.com/tech/react19-rsc-04-use-api">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 29 Jan 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>react</category>
      <category>nextjs</category>
      <category>performance</category>
      <category>frontend</category>
      <category>typescript</category>
    </item>
    <item>
      <title><![CDATA[6장: Kubernetes 배포 실전 - GPU 노드와 모델 서빙 배포]]></title>
      <link>https://archive.kreathlab.com/tech/ai-deployment-06-kubernetes-deployment</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-deployment-06-kubernetes-deployment</guid>
      <description><![CDATA[Kubernetes에서 GPU 기반 AI 서비스를 배포하는 실전 과정을 다루며, 프로브 설정, 리소스 관리, 무중단 배포 전략을 구현합니다.]]></description>
      <content:encoded><![CDATA[<p>Kubernetes에서 GPU 기반 AI 서비스를 배포하는 실전 과정을 다루며, 프로브 설정, 리소스 관리, 무중단 배포 전략을 구현합니다.</p><p><strong>AI / ML</strong> / 18분 읽기 / 시리즈: ai-deployment</p><p><a href="https://archive.kreathlab.com/tech/ai-deployment-06-kubernetes-deployment">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 28 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>mlops</category>
      <category>kubernetes</category>
      <category>infrastructure</category>
      <category>performance</category>
    </item>
    <item>
      <title><![CDATA[8장: 모델 레지스트리와 버전 관리]]></title>
      <link>https://archive.kreathlab.com/tech/fine-tuning-08-model-registry</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/fine-tuning-08-model-registry</guid>
      <description><![CDATA[파인튜닝된 모델을 체계적으로 관리하기 위한 모델 레지스트리 구축, 버전 관리, 메타데이터 추적, 아티팩트 저장 전략을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>파인튜닝된 모델을 체계적으로 관리하기 위한 모델 레지스트리 구축, 버전 관리, 메타데이터 추적, 아티팩트 저장 전략을 다룹니다.</p><p><strong>AI / ML</strong> / 15분 읽기 / 시리즈: fine-tuning</p><p><a href="https://archive.kreathlab.com/tech/fine-tuning-08-model-registry">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 28 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>training</category>
      <category>mlops</category>
      <category>data-engineering</category>
    </item>
    <item>
      <title><![CDATA[5장: 데이터 페칭과 캐싱 전략의 대전환]]></title>
      <link>https://archive.kreathlab.com/tech/nextjs-app-router-05-data-fetching</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/nextjs-app-router-05-data-fetching</guid>
      <description><![CDATA[Next.js 14에서 16까지 캐싱 전략이 어떻게 변화했는지 살펴봅니다. 네 가지 캐싱 레이어, fetch() 기본값 변경, 재검증 전략을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Next.js 14에서 16까지 캐싱 전략이 어떻게 변화했는지 살펴봅니다. 네 가지 캐싱 레이어, fetch() 기본값 변경, 재검증 전략을 다룹니다.</p><p><strong>웹 개발</strong> / 20분 읽기 / 시리즈: nextjs-app-router</p><p><a href="https://archive.kreathlab.com/tech/nextjs-app-router-05-data-fetching">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 28 Jan 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>nextjs</category>
      <category>react</category>
      <category>typescript</category>
      <category>frontend</category>
    </item>
    <item>
      <title><![CDATA[9장: 프롬프트 테스트와 평가 자동화]]></title>
      <link>https://archive.kreathlab.com/tech/prompt-engineering-09-testing</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/prompt-engineering-09-testing</guid>
      <description><![CDATA[프롬프트의 품질을 정량적으로 측정하고 회귀를 방지하는 체계적인 테스트 전략과 자동화 도구를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>프롬프트의 품질을 정량적으로 측정하고 회귀를 방지하는 체계적인 테스트 전략과 자동화 도구를 다룹니다.</p><p><strong>AI / ML</strong> / 18분 읽기 / 시리즈: prompt-engineering</p><p><a href="https://archive.kreathlab.com/tech/prompt-engineering-09-testing">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 28 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>prompt-engineering</category>
      <category>structured-output</category>
      <category>training</category>
    </item>
    <item>
      <title><![CDATA[7장: Free-threaded Python - GIL 제거의 시작]]></title>
      <link>https://archive.kreathlab.com/tech/python-trends-07-free-threaded</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/python-trends-07-free-threaded</guid>
      <description><![CDATA[Python 3.13의 free-threaded 모드를 심층 분석합니다. GIL의 역사와 문제점, PEP 703의 설계, free-threaded 빌드의 설치와 실전 멀티스레드 성능을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Python 3.13의 free-threaded 모드를 심층 분석합니다. GIL의 역사와 문제점, PEP 703의 설계, free-threaded 빌드의 설치와 실전 멀티스레드 성능을 다룹니다.</p><p><strong>프로그래밍</strong> / 16분 읽기 / 시리즈: python-trends</p><p><a href="https://archive.kreathlab.com/tech/python-trends-07-free-threaded">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 28 Jan 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>python</category>
      <category>typescript</category>
      <category>performance</category>
      <category>devtools</category>
      <category>concurrency</category>
    </item>
    <item>
      <title><![CDATA[4장: using 선언과 명시적 리소스 관리]]></title>
      <link>https://archive.kreathlab.com/tech/typescript-deepdive-04-using-declarations</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/typescript-deepdive-04-using-declarations</guid>
      <description><![CDATA[TypeScript 5.2에서 도입된 using 선언과 Symbol.dispose를 활용한 명시적 리소스 관리 패턴을 실전 예제와 함께 심층 분석합니다.]]></description>
      <content:encoded><![CDATA[<p>TypeScript 5.2에서 도입된 using 선언과 Symbol.dispose를 활용한 명시적 리소스 관리 패턴을 실전 예제와 함께 심층 분석합니다.</p><p><strong>프로그래밍</strong> / 12분 읽기 / 시리즈: typescript-deepdive</p><p><a href="https://archive.kreathlab.com/tech/typescript-deepdive-04-using-declarations">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 28 Jan 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>typescript</category>
      <category>performance</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[6장: 멀티 에이전트 패턴 - 협업과 조율의 아키텍처]]></title>
      <link>https://archive.kreathlab.com/tech/ai-agent-patterns-06-multi-agent</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-agent-patterns-06-multi-agent</guid>
      <description><![CDATA[여러 전문화된 에이전트가 협업하는 멀티 에이전트 시스템의 설계 패턴, 감독자/토론/파이프라인 아키텍처를 코드와 함께 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>여러 전문화된 에이전트가 협업하는 멀티 에이전트 시스템의 설계 패턴, 감독자/토론/파이프라인 아키텍처를 코드와 함께 다룹니다.</p><p><strong>AI / ML</strong> / 15분 읽기 / 시리즈: ai-agent-patterns</p><p><a href="https://archive.kreathlab.com/tech/ai-agent-patterns-06-multi-agent">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 27 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai-agent</category>
      <category>llm</category>
      <category>architecture</category>
      <category>orchestration</category>
    </item>
    <item>
      <title><![CDATA[5장: AI 기반 문서화 자동화]]></title>
      <link>https://archive.kreathlab.com/tech/ai-dev-workflow-05-documentation</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-dev-workflow-05-documentation</guid>
      <description><![CDATA[코드 변경에 따라 API 문서, README, 변경 로그를 AI로 자동 갱신하는 시스템을 구축하고, 문서와 코드의 동기화를 유지하는 전략을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>코드 변경에 따라 API 문서, README, 변경 로그를 AI로 자동 갱신하는 시스템을 구축하고, 문서와 코드의 동기화를 유지하는 전략을 다룹니다.</p><p><strong>AI / ML</strong> / 19분 읽기 / 시리즈: ai-dev-workflow</p><p><a href="https://archive.kreathlab.com/tech/ai-dev-workflow-05-documentation">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 27 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>devtools</category>
      <category>automation</category>
      <category>code-quality</category>
      <category>devops</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[7장: 프로덕션 로깅과 관찰 가능성]]></title>
      <link>https://archive.kreathlab.com/tech/llm-evaluation-07-logging-observability</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/llm-evaluation-07-logging-observability</guid>
      <description><![CDATA[LLM 애플리케이션의 프로덕션 환경에서 구조화된 로깅, 분산 트레이싱, 관찰 가능성을 구축하는 방법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>LLM 애플리케이션의 프로덕션 환경에서 구조화된 로깅, 분산 트레이싱, 관찰 가능성을 구축하는 방법을 다룹니다.</p><p><strong>AI / ML</strong> / 14분 읽기 / 시리즈: llm-evaluation</p><p><a href="https://archive.kreathlab.com/tech/llm-evaluation-07-logging-observability">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 27 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>evaluation</category>
      <category>monitoring</category>
      <category>observability</category>
      <category>testing</category>
    </item>
    <item>
      <title><![CDATA[4장: 서버 프리미티브 - 도구, 리소스, 프롬프트]]></title>
      <link>https://archive.kreathlab.com/tech/mcp-guide-04-server-primitives</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/mcp-guide-04-server-primitives</guid>
      <description><![CDATA[MCP 서버가 제공하는 세 가지 핵심 프리미티브의 스키마 정의, 구현 패턴, 실전 활용 사례를 상세히 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>MCP 서버가 제공하는 세 가지 핵심 프리미티브의 스키마 정의, 구현 패턴, 실전 활용 사례를 상세히 다룹니다.</p><p><strong>AI / ML</strong> / 24분 읽기 / 시리즈: mcp-guide</p><p><a href="https://archive.kreathlab.com/tech/mcp-guide-04-server-primitives">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 27 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>mcp</category>
      <category>typescript</category>
      <category>python</category>
    </item>
    <item>
      <title><![CDATA[8장: RAG 평가 프레임워크와 메트릭]]></title>
      <link>https://archive.kreathlab.com/tech/rag-system-08-evaluation</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/rag-system-08-evaluation</guid>
      <description><![CDATA[RAGAS, 충실도, 컨텍스트 정밀도 등 RAG 시스템의 품질을 객관적으로 측정하는 평가 프레임워크와 핵심 메트릭을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>RAGAS, 충실도, 컨텍스트 정밀도 등 RAG 시스템의 품질을 객관적으로 측정하는 평가 프레임워크와 핵심 메트릭을 다룹니다.</p><p><strong>AI / ML</strong> / 18분 읽기 / 시리즈: rag-system</p><p><a href="https://archive.kreathlab.com/tech/rag-system-08-evaluation">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 27 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>rag</category>
      <category>vector-database</category>
      <category>embedding</category>
      <category>retrieval</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[3장: Server Actions로 서버-클라이언트 통합하기]]></title>
      <link>https://archive.kreathlab.com/tech/react19-rsc-03-server-actions</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/react19-rsc-03-server-actions</guid>
      <description><![CDATA[Server Actions의 동작 원리, 폼 처리 패턴, 데이터 뮤테이션, 에러 핸들링, 보안 고려사항을 실전 코드와 함께 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Server Actions의 동작 원리, 폼 처리 패턴, 데이터 뮤테이션, 에러 핸들링, 보안 고려사항을 실전 코드와 함께 다룹니다.</p><p><strong>웹 개발</strong> / 22분 읽기 / 시리즈: react19-rsc</p><p><a href="https://archive.kreathlab.com/tech/react19-rsc-03-server-actions">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 27 Jan 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>react</category>
      <category>nextjs</category>
      <category>performance</category>
      <category>frontend</category>
      <category>typescript</category>
    </item>
    <item>
      <title><![CDATA[5장: Kubernetes 기초 - AI 워크로드를 위한 클러스터 설계]]></title>
      <link>https://archive.kreathlab.com/tech/ai-deployment-05-kubernetes-basics</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-deployment-05-kubernetes-basics</guid>
      <description><![CDATA[Kubernetes의 핵심 개념을 AI 워크로드 관점에서 설명하고, GPU 노드 구성과 AI 서비스에 적합한 클러스터 아키텍처를 설계합니다.]]></description>
      <content:encoded><![CDATA[<p>Kubernetes의 핵심 개념을 AI 워크로드 관점에서 설명하고, GPU 노드 구성과 AI 서비스에 적합한 클러스터 아키텍처를 설계합니다.</p><p><strong>AI / ML</strong> / 14분 읽기 / 시리즈: ai-deployment</p><p><a href="https://archive.kreathlab.com/tech/ai-deployment-05-kubernetes-basics">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 26 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>mlops</category>
      <category>kubernetes</category>
      <category>infrastructure</category>
      <category>performance</category>
    </item>
    <item>
      <title><![CDATA[7장: 파인튜닝 모델 평가와 벤치마킹]]></title>
      <link>https://archive.kreathlab.com/tech/fine-tuning-07-evaluation</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/fine-tuning-07-evaluation</guid>
      <description><![CDATA[파인튜닝된 모델의 성능을 자동 메트릭, LLM 평가, 인간 평가를 통해 다각적으로 측정하고 벤치마킹하는 체계적인 방법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>파인튜닝된 모델의 성능을 자동 메트릭, LLM 평가, 인간 평가를 통해 다각적으로 측정하고 벤치마킹하는 체계적인 방법을 다룹니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: fine-tuning</p><p><a href="https://archive.kreathlab.com/tech/fine-tuning-07-evaluation">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 26 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>training</category>
      <category>mlops</category>
      <category>data-engineering</category>
    </item>
    <item>
      <title><![CDATA[4장: 인터셉팅 라우트와 모달 패턴]]></title>
      <link>https://archive.kreathlab.com/tech/nextjs-app-router-04-intercepting-routes</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/nextjs-app-router-04-intercepting-routes</guid>
      <description><![CDATA[Next.js의 인터셉팅 라우트로 모달, 사진 갤러리, 미리보기 패턴을 구현합니다. 병렬 라우트와의 조합, 딥 링킹, 공유 가능한 URL을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Next.js의 인터셉팅 라우트로 모달, 사진 갤러리, 미리보기 패턴을 구현합니다. 병렬 라우트와의 조합, 딥 링킹, 공유 가능한 URL을 다룹니다.</p><p><strong>웹 개발</strong> / 12분 읽기 / 시리즈: nextjs-app-router</p><p><a href="https://archive.kreathlab.com/tech/nextjs-app-router-04-intercepting-routes">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 26 Jan 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>nextjs</category>
      <category>react</category>
      <category>typescript</category>
      <category>frontend</category>
    </item>
    <item>
      <title><![CDATA[8장: 고급 기법 - 메타 프롬프팅, 프롬프트 체이닝, 자기 성찰]]></title>
      <link>https://archive.kreathlab.com/tech/prompt-engineering-08-advanced-techniques</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/prompt-engineering-08-advanced-techniques</guid>
      <description><![CDATA[메타 프롬프팅, 프롬프트 체이닝, 자기 성찰, Tree-of-Thought 등 복잡한 작업을 해결하는 고급 프롬프트 엔지니어링 기법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>메타 프롬프팅, 프롬프트 체이닝, 자기 성찰, Tree-of-Thought 등 복잡한 작업을 해결하는 고급 프롬프트 엔지니어링 기법을 다룹니다.</p><p><strong>AI / ML</strong> / 22분 읽기 / 시리즈: prompt-engineering</p><p><a href="https://archive.kreathlab.com/tech/prompt-engineering-08-advanced-techniques">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 26 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>prompt-engineering</category>
      <category>structured-output</category>
      <category>training</category>
    </item>
    <item>
      <title><![CDATA[6장: CPython 성능 향상의 메커니즘]]></title>
      <link>https://archive.kreathlab.com/tech/python-trends-06-performance</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/python-trends-06-performance</guid>
      <description><![CDATA[Python 3.12의 성능 향상 원리를 분석합니다. 특수화 적응 인터프리터, 컴프리헨션 인라인화, immortal objects, asyncio 최적화 등 CPython 내부를 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Python 3.12의 성능 향상 원리를 분석합니다. 특수화 적응 인터프리터, 컴프리헨션 인라인화, immortal objects, asyncio 최적화 등 CPython 내부를 다룹니다.</p><p><strong>프로그래밍</strong> / 16분 읽기 / 시리즈: python-trends</p><p><a href="https://archive.kreathlab.com/tech/python-trends-06-performance">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 26 Jan 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>python</category>
      <category>typescript</category>
      <category>performance</category>
      <category>devtools</category>
      <category>concurrency</category>
    </item>
    <item>
      <title><![CDATA[3장: TC39 표준 데코레이터 완벽 이해]]></title>
      <link>https://archive.kreathlab.com/tech/typescript-deepdive-03-decorators</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/typescript-deepdive-03-decorators</guid>
      <description><![CDATA[TypeScript 5.0에서 도입된 TC39 Stage 3 데코레이터의 원리, API 구조, 실전 패턴을 다루고, 기존 실험적 데코레이터와의 차이를 분석합니다.]]></description>
      <content:encoded><![CDATA[<p>TypeScript 5.0에서 도입된 TC39 Stage 3 데코레이터의 원리, API 구조, 실전 패턴을 다루고, 기존 실험적 데코레이터와의 차이를 분석합니다.</p><p><strong>프로그래밍</strong> / 14분 읽기 / 시리즈: typescript-deepdive</p><p><a href="https://archive.kreathlab.com/tech/typescript-deepdive-03-decorators">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 26 Jan 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>typescript</category>
      <category>performance</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[5장: 계획 수립 패턴 - 복잡한 작업의 분해와 실행]]></title>
      <link>https://archive.kreathlab.com/tech/ai-agent-patterns-05-planning</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-agent-patterns-05-planning</guid>
      <description><![CDATA[Plan-and-Execute 아키텍처의 원리와 구현, 적응적 재계획 전략, 그리고 계획 수립 패턴이 에이전트 성능에 미치는 영향을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Plan-and-Execute 아키텍처의 원리와 구현, 적응적 재계획 전략, 그리고 계획 수립 패턴이 에이전트 성능에 미치는 영향을 다룹니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: ai-agent-patterns</p><p><a href="https://archive.kreathlab.com/tech/ai-agent-patterns-05-planning">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 25 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai-agent</category>
      <category>llm</category>
      <category>architecture</category>
      <category>orchestration</category>
    </item>
    <item>
      <title><![CDATA[4장: AI 기반 테스트 자동 생성]]></title>
      <link>https://archive.kreathlab.com/tech/ai-dev-workflow-04-test-generation</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-dev-workflow-04-test-generation</guid>
      <description><![CDATA[코드 변경을 분석하여 단위 테스트와 통합 테스트를 자동으로 생성하는 시스템을 구축하고, 테스트 품질을 검증하는 방법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>코드 변경을 분석하여 단위 테스트와 통합 테스트를 자동으로 생성하는 시스템을 구축하고, 테스트 품질을 검증하는 방법을 다룹니다.</p><p><strong>AI / ML</strong> / 18분 읽기 / 시리즈: ai-dev-workflow</p><p><a href="https://archive.kreathlab.com/tech/ai-dev-workflow-04-test-generation">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 25 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>devtools</category>
      <category>automation</category>
      <category>code-quality</category>
      <category>devops</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[6장: A/B 테스트와 온라인 실험]]></title>
      <link>https://archive.kreathlab.com/tech/llm-evaluation-06-ab-testing</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/llm-evaluation-06-ab-testing</guid>
      <description><![CDATA[LLM 애플리케이션에서 A/B 테스트를 설계하고 실행하는 방법, 통계적 유의성 판단, 실험 결과 해석을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>LLM 애플리케이션에서 A/B 테스트를 설계하고 실행하는 방법, 통계적 유의성 판단, 실험 결과 해석을 다룹니다.</p><p><strong>AI / ML</strong> / 15분 읽기 / 시리즈: llm-evaluation</p><p><a href="https://archive.kreathlab.com/tech/llm-evaluation-06-ab-testing">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 25 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>evaluation</category>
      <category>monitoring</category>
      <category>observability</category>
      <category>testing</category>
    </item>
    <item>
      <title><![CDATA[3장: 전송 계층 - stdio와 Streamable HTTP]]></title>
      <link>https://archive.kreathlab.com/tech/mcp-guide-03-transport-layer</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/mcp-guide-03-transport-layer</guid>
      <description><![CDATA[MCP의 두 가지 핵심 전송 방식인 stdio와 Streamable HTTP의 동작 원리, 장단점, 선택 기준을 상세히 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>MCP의 두 가지 핵심 전송 방식인 stdio와 Streamable HTTP의 동작 원리, 장단점, 선택 기준을 상세히 다룹니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: mcp-guide</p><p><a href="https://archive.kreathlab.com/tech/mcp-guide-03-transport-layer">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 25 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>mcp</category>
      <category>typescript</category>
      <category>python</category>
    </item>
    <item>
      <title><![CDATA[7장: 리랭킹으로 검색 정밀도 높이기]]></title>
      <link>https://archive.kreathlab.com/tech/rag-system-07-reranking</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/rag-system-07-reranking</guid>
      <description><![CDATA[Cross-Encoder 리랭킹의 원리, Cohere Rerank API, 오픈소스 리랭커 비교, 그리고 프로덕션 환경에서의 효과적인 리랭킹 전략을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Cross-Encoder 리랭킹의 원리, Cohere Rerank API, 오픈소스 리랭커 비교, 그리고 프로덕션 환경에서의 효과적인 리랭킹 전략을 다룹니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: rag-system</p><p><a href="https://archive.kreathlab.com/tech/rag-system-07-reranking">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 25 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>rag</category>
      <category>vector-database</category>
      <category>embedding</category>
      <category>retrieval</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[2장: React Server Components 아키텍처 심층 분석]]></title>
      <link>https://archive.kreathlab.com/tech/react19-rsc-02-server-components</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/react19-rsc-02-server-components</guid>
      <description><![CDATA[React Server Components의 동작 원리, 직렬화 프로토콜, 번들 전략, 합성 규칙을 심층적으로 분석합니다. 서버와 클라이언트의 경계를 이해합니다.]]></description>
      <content:encoded><![CDATA[<p>React Server Components의 동작 원리, 직렬화 프로토콜, 번들 전략, 합성 규칙을 심층적으로 분석합니다. 서버와 클라이언트의 경계를 이해합니다.</p><p><strong>웹 개발</strong> / 23분 읽기 / 시리즈: react19-rsc</p><p><a href="https://archive.kreathlab.com/tech/react19-rsc-02-server-components">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 25 Jan 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>react</category>
      <category>nextjs</category>
      <category>performance</category>
      <category>frontend</category>
      <category>typescript</category>
    </item>
    <item>
      <title><![CDATA[4장: 컨테이너화 - Docker로 AI 서비스 패키징]]></title>
      <link>https://archive.kreathlab.com/tech/ai-deployment-04-containerization</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-deployment-04-containerization</guid>
      <description><![CDATA[GPU 지원 Docker 컨테이너로 AI 서비스를 패키징하는 방법을 다루며, NVIDIA Container Toolkit 설정부터 멀티 스테이지 빌드까지 실전 기법을 소개합니다.]]></description>
      <content:encoded><![CDATA[<p>GPU 지원 Docker 컨테이너로 AI 서비스를 패키징하는 방법을 다루며, NVIDIA Container Toolkit 설정부터 멀티 스테이지 빌드까지 실전 기법을 소개합니다.</p><p><strong>AI / ML</strong> / 15분 읽기 / 시리즈: ai-deployment</p><p><a href="https://archive.kreathlab.com/tech/ai-deployment-04-containerization">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 24 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>mlops</category>
      <category>kubernetes</category>
      <category>infrastructure</category>
      <category>performance</category>
    </item>
    <item>
      <title><![CDATA[6장: 학습 파이프라인 구축과 하이퍼파라미터 최적화]]></title>
      <link>https://archive.kreathlab.com/tech/fine-tuning-06-training-pipeline</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/fine-tuning-06-training-pipeline</guid>
      <description><![CDATA[파인튜닝 학습 파이프라인의 전체 구조를 설계하고, 학습률, 배치 크기, 스케줄링 등 핵심 하이퍼파라미터를 최적화하는 전략을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>파인튜닝 학습 파이프라인의 전체 구조를 설계하고, 학습률, 배치 크기, 스케줄링 등 핵심 하이퍼파라미터를 최적화하는 전략을 다룹니다.</p><p><strong>AI / ML</strong> / 18분 읽기 / 시리즈: fine-tuning</p><p><a href="https://archive.kreathlab.com/tech/fine-tuning-06-training-pipeline">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 24 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>training</category>
      <category>mlops</category>
      <category>data-engineering</category>
    </item>
    <item>
      <title><![CDATA[3장: 동적 라우팅과 병렬 라우트]]></title>
      <link>https://archive.kreathlab.com/tech/nextjs-app-router-03-parallel-routes</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/nextjs-app-router-03-parallel-routes</guid>
      <description><![CDATA[Next.js의 동적 세그먼트, catch-all 라우트, 병렬 라우트(@slot)의 동작 원리와 대시보드, 조건부 렌더링 등 실전 패턴을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Next.js의 동적 세그먼트, catch-all 라우트, 병렬 라우트(@slot)의 동작 원리와 대시보드, 조건부 렌더링 등 실전 패턴을 다룹니다.</p><p><strong>웹 개발</strong> / 10분 읽기 / 시리즈: nextjs-app-router</p><p><a href="https://archive.kreathlab.com/tech/nextjs-app-router-03-parallel-routes">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 24 Jan 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>nextjs</category>
      <category>react</category>
      <category>typescript</category>
      <category>frontend</category>
    </item>
    <item>
      <title><![CDATA[7장: 시스템 프롬프트 설계 패턴]]></title>
      <link>https://archive.kreathlab.com/tech/prompt-engineering-07-system-prompt</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/prompt-engineering-07-system-prompt</guid>
      <description><![CDATA[프로덕션 환경에서 일관된 모델 행동을 보장하는 시스템 프롬프트의 구조, 설계 원칙, 그리고 실전 패턴을 체계적으로 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>프로덕션 환경에서 일관된 모델 행동을 보장하는 시스템 프롬프트의 구조, 설계 원칙, 그리고 실전 패턴을 체계적으로 다룹니다.</p><p><strong>AI / ML</strong> / 20분 읽기 / 시리즈: prompt-engineering</p><p><a href="https://archive.kreathlab.com/tech/prompt-engineering-07-system-prompt">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 24 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>prompt-engineering</category>
      <category>structured-output</category>
      <category>training</category>
    </item>
    <item>
      <title><![CDATA[5장: PyREPL - 새로운 대화형 인터프리터]]></title>
      <link>https://archive.kreathlab.com/tech/python-trends-05-pyrepl</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/python-trends-05-pyrepl</guid>
      <description><![CDATA[Python 3.13에서 도입된 PyREPL의 구문 강조, 멀티라인 편집, 자동완성, 히스토리 관리 등 현대적 REPL 기능을 실전 예시와 함께 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Python 3.13에서 도입된 PyREPL의 구문 강조, 멀티라인 편집, 자동완성, 히스토리 관리 등 현대적 REPL 기능을 실전 예시와 함께 다룹니다.</p><p><strong>프로그래밍</strong> / 15분 읽기 / 시리즈: python-trends</p><p><a href="https://archive.kreathlab.com/tech/python-trends-05-pyrepl">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 24 Jan 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>python</category>
      <category>typescript</category>
      <category>performance</category>
      <category>devtools</category>
      <category>concurrency</category>
    </item>
    <item>
      <title><![CDATA[2장: const 타입 파라미터와 satisfies 연산자]]></title>
      <link>https://archive.kreathlab.com/tech/typescript-deepdive-02-const-satisfies</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/typescript-deepdive-02-const-satisfies</guid>
      <description><![CDATA[TypeScript의 타입 추론을 정밀하게 제어하는 두 가지 핵심 도구인 const 타입 파라미터와 satisfies 연산자의 원리, 차이점, 실전 활용 패턴을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>TypeScript의 타입 추론을 정밀하게 제어하는 두 가지 핵심 도구인 const 타입 파라미터와 satisfies 연산자의 원리, 차이점, 실전 활용 패턴을 다룹니다.</p><p><strong>프로그래밍</strong> / 15분 읽기 / 시리즈: typescript-deepdive</p><p><a href="https://archive.kreathlab.com/tech/typescript-deepdive-02-const-satisfies">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 24 Jan 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>typescript</category>
      <category>performance</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[4장: 리플렉션 패턴 - 자기 평가와 반복 개선]]></title>
      <link>https://archive.kreathlab.com/tech/ai-agent-patterns-04-reflection</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-agent-patterns-04-reflection</guid>
      <description><![CDATA[에이전트가 자신의 출력을 평가하고 반복적으로 개선하는 리플렉션 패턴의 원리, 구현 방법, 실전 활용 전략을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>에이전트가 자신의 출력을 평가하고 반복적으로 개선하는 리플렉션 패턴의 원리, 구현 방법, 실전 활용 전략을 다룹니다.</p><p><strong>AI / ML</strong> / 18분 읽기 / 시리즈: ai-agent-patterns</p><p><a href="https://archive.kreathlab.com/tech/ai-agent-patterns-04-reflection">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 23 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai-agent</category>
      <category>llm</category>
      <category>architecture</category>
      <category>orchestration</category>
    </item>
    <item>
      <title><![CDATA[3장: AI 코드 리뷰 실전 구축 - GitHub Actions 통합]]></title>
      <link>https://archive.kreathlab.com/tech/ai-dev-workflow-03-code-review-implementation</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-dev-workflow-03-code-review-implementation</guid>
      <description><![CDATA[GitHub Actions를 활용하여 PR에 자동으로 AI 코드 리뷰를 수행하는 시스템을 직접 구축하고, 실전에서 활용 가능한 수준으로 완성합니다.]]></description>
      <content:encoded><![CDATA[<p>GitHub Actions를 활용하여 PR에 자동으로 AI 코드 리뷰를 수행하는 시스템을 직접 구축하고, 실전에서 활용 가능한 수준으로 완성합니다.</p><p><strong>AI / ML</strong> / 18분 읽기 / 시리즈: ai-dev-workflow</p><p><a href="https://archive.kreathlab.com/tech/ai-dev-workflow-03-code-review-implementation">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 23 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>devtools</category>
      <category>automation</category>
      <category>code-quality</category>
      <category>devops</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[5장: 인간 평가와 어노테이션 설계]]></title>
      <link>https://archive.kreathlab.com/tech/llm-evaluation-05-human-evaluation</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/llm-evaluation-05-human-evaluation</guid>
      <description><![CDATA[LLM 평가에서 인간 평가의 역할, 어노테이션 가이드라인 설계, 평가자 간 일치도 관리 방법을 체계적으로 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>LLM 평가에서 인간 평가의 역할, 어노테이션 가이드라인 설계, 평가자 간 일치도 관리 방법을 체계적으로 다룹니다.</p><p><strong>AI / ML</strong> / 20분 읽기 / 시리즈: llm-evaluation</p><p><a href="https://archive.kreathlab.com/tech/llm-evaluation-05-human-evaluation">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 23 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>evaluation</category>
      <category>monitoring</category>
      <category>observability</category>
      <category>testing</category>
    </item>
    <item>
      <title><![CDATA[2장: 프로토콜 아키텍처 심층 분석]]></title>
      <link>https://archive.kreathlab.com/tech/mcp-guide-02-protocol-deep-dive</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/mcp-guide-02-protocol-deep-dive</guid>
      <description><![CDATA[MCP의 JSON-RPC 2.0 기반 메시지 형식, 연결 생명주기, 능력 협상 메커니즘을 상세히 분석합니다.]]></description>
      <content:encoded><![CDATA[<p>MCP의 JSON-RPC 2.0 기반 메시지 형식, 연결 생명주기, 능력 협상 메커니즘을 상세히 분석합니다.</p><p><strong>AI / ML</strong> / 20분 읽기 / 시리즈: mcp-guide</p><p><a href="https://archive.kreathlab.com/tech/mcp-guide-02-protocol-deep-dive">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 23 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>mcp</category>
      <category>typescript</category>
      <category>python</category>
    </item>
    <item>
      <title><![CDATA[6장: 하이브리드 검색 - BM25와 시맨틱 검색의 결합]]></title>
      <link>https://archive.kreathlab.com/tech/rag-system-06-hybrid-search</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/rag-system-06-hybrid-search</guid>
      <description><![CDATA[키워드 기반 BM25와 벡터 기반 시맨틱 검색을 결합한 하이브리드 검색의 원리, 구현 방법, 그리고 Reciprocal Rank Fusion 전략을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>키워드 기반 BM25와 벡터 기반 시맨틱 검색을 결합한 하이브리드 검색의 원리, 구현 방법, 그리고 Reciprocal Rank Fusion 전략을 다룹니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: rag-system</p><p><a href="https://archive.kreathlab.com/tech/rag-system-06-hybrid-search">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 23 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>rag</category>
      <category>vector-database</category>
      <category>embedding</category>
      <category>retrieval</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[1장: React 19의 등장과 새로운 패러다임]]></title>
      <link>https://archive.kreathlab.com/tech/react19-rsc-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/react19-rsc-01-introduction</guid>
      <description><![CDATA[React 19가 가져온 근본적인 변화를 살펴봅니다. Actions, Server Components, 새로운 훅, React Compiler까지 React의 새로운 패러다임을 이해합니다.]]></description>
      <content:encoded><![CDATA[<p>React 19가 가져온 근본적인 변화를 살펴봅니다. Actions, Server Components, 새로운 훅, React Compiler까지 React의 새로운 패러다임을 이해합니다.</p><p><strong>웹 개발</strong> / 16분 읽기 / 시리즈: react19-rsc</p><p><a href="https://archive.kreathlab.com/tech/react19-rsc-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 23 Jan 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>react</category>
      <category>nextjs</category>
      <category>performance</category>
      <category>frontend</category>
      <category>typescript</category>
    </item>
    <item>
      <title><![CDATA[3장: 모델 최적화 - 양자화, 배칭, KV 캐시 전략]]></title>
      <link>https://archive.kreathlab.com/tech/ai-deployment-03-model-optimization</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-deployment-03-model-optimization</guid>
      <description><![CDATA[LLM 추론 성능을 극대화하기 위한 양자화 기법, 배칭 전략, KV 캐시 튜닝 방법을 실전 예제와 함께 체계적으로 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>LLM 추론 성능을 극대화하기 위한 양자화 기법, 배칭 전략, KV 캐시 튜닝 방법을 실전 예제와 함께 체계적으로 다룹니다.</p><p><strong>AI / ML</strong> / 22분 읽기 / 시리즈: ai-deployment</p><p><a href="https://archive.kreathlab.com/tech/ai-deployment-03-model-optimization">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 22 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>mlops</category>
      <category>kubernetes</category>
      <category>infrastructure</category>
      <category>performance</category>
    </item>
    <item>
      <title><![CDATA[5장: QLoRA로 소비자 GPU에서 파인튜닝하기]]></title>
      <link>https://archive.kreathlab.com/tech/fine-tuning-05-qlora</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/fine-tuning-05-qlora</guid>
      <description><![CDATA[4비트 양자화와 LoRA를 결합한 QLoRA의 원리를 이해하고, 단일 소비자 GPU에서 대규모 모델을 파인튜닝하는 실전 방법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>4비트 양자화와 LoRA를 결합한 QLoRA의 원리를 이해하고, 단일 소비자 GPU에서 대규모 모델을 파인튜닝하는 실전 방법을 다룹니다.</p><p><strong>AI / ML</strong> / 14분 읽기 / 시리즈: fine-tuning</p><p><a href="https://archive.kreathlab.com/tech/fine-tuning-05-qlora">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 22 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>training</category>
      <category>mlops</category>
      <category>data-engineering</category>
    </item>
    <item>
      <title><![CDATA[2장: App Router 아키텍처 - 레이아웃, 템플릿, 라우트 그룹]]></title>
      <link>https://archive.kreathlab.com/tech/nextjs-app-router-02-architecture</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/nextjs-app-router-02-architecture</guid>
      <description><![CDATA[Next.js App Router의 파일 기반 라우팅, 레이아웃 계층, 템플릿, 라우트 그룹, 에러/로딩 경계의 동작 원리와 실전 패턴을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Next.js App Router의 파일 기반 라우팅, 레이아웃 계층, 템플릿, 라우트 그룹, 에러/로딩 경계의 동작 원리와 실전 패턴을 다룹니다.</p><p><strong>웹 개발</strong> / 15분 읽기 / 시리즈: nextjs-app-router</p><p><a href="https://archive.kreathlab.com/tech/nextjs-app-router-02-architecture">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 22 Jan 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>nextjs</category>
      <category>react</category>
      <category>typescript</category>
      <category>frontend</category>
    </item>
    <item>
      <title><![CDATA[6장: 구조화된 출력 - JSON Schema와 타입 안전 응답]]></title>
      <link>https://archive.kreathlab.com/tech/prompt-engineering-06-structured-output</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/prompt-engineering-06-structured-output</guid>
      <description><![CDATA[LLM이 JSON Schema를 따르는 구조화된 응답을 생성하도록 설계하는 방법과 프로덕션 시스템 통합 전략을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>LLM이 JSON Schema를 따르는 구조화된 응답을 생성하도록 설계하는 방법과 프로덕션 시스템 통합 전략을 다룹니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: prompt-engineering</p><p><a href="https://archive.kreathlab.com/tech/prompt-engineering-06-structured-output">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 22 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>prompt-engineering</category>
      <category>structured-output</category>
      <category>training</category>
    </item>
    <item>
      <title><![CDATA[4장: 유연한 f-string과 개선된 에러 메시지]]></title>
      <link>https://archive.kreathlab.com/tech/python-trends-04-fstring-errors</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/python-trends-04-fstring-errors</guid>
      <description><![CDATA[Python 3.12의 PEP 701 유연한 f-string 파싱과 3.12~3.13의 에러 메시지 개선을 다룹니다. 컬러 트레이스백, 제안 기반 에러 등 디버깅 경험 향상을 살펴봅니다.]]></description>
      <content:encoded><![CDATA[<p>Python 3.12의 PEP 701 유연한 f-string 파싱과 3.12~3.13의 에러 메시지 개선을 다룹니다. 컬러 트레이스백, 제안 기반 에러 등 디버깅 경험 향상을 살펴봅니다.</p><p><strong>프로그래밍</strong> / 13분 읽기 / 시리즈: python-trends</p><p><a href="https://archive.kreathlab.com/tech/python-trends-04-fstring-errors">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 22 Jan 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>python</category>
      <category>typescript</category>
      <category>performance</category>
      <category>devtools</category>
      <category>concurrency</category>
    </item>
    <item>
      <title><![CDATA[1장: TypeScript 5.x의 진화 — 5.0부터 5.8까지]]></title>
      <link>https://archive.kreathlab.com/tech/typescript-deepdive-01-evolution</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/typescript-deepdive-01-evolution</guid>
      <description><![CDATA[TypeScript 5.x 시리즈의 주요 변경사항을 버전별로 정리하고, 타입 시스템의 진화 방향과 개발자 경험 개선을 조망합니다.]]></description>
      <content:encoded><![CDATA[<p>TypeScript 5.x 시리즈의 주요 변경사항을 버전별로 정리하고, 타입 시스템의 진화 방향과 개발자 경험 개선을 조망합니다.</p><p><strong>프로그래밍</strong> / 14분 읽기 / 시리즈: typescript-deepdive</p><p><a href="https://archive.kreathlab.com/tech/typescript-deepdive-01-evolution">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 22 Jan 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>typescript</category>
      <category>performance</category>
      <category>devtools</category>
    </item>
    <item>
      <title><![CDATA[3장: 도구 사용 패턴 - 에이전트의 손과 발]]></title>
      <link>https://archive.kreathlab.com/tech/ai-agent-patterns-03-tool-use</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-agent-patterns-03-tool-use</guid>
      <description><![CDATA[AI 에이전트의 도구 정의, 호출, 결과 통합의 전 과정을 다루고, 효과적인 도구 스키마 설계와 복합 도구 조합 전략을 살펴봅니다.]]></description>
      <content:encoded><![CDATA[<p>AI 에이전트의 도구 정의, 호출, 결과 통합의 전 과정을 다루고, 효과적인 도구 스키마 설계와 복합 도구 조합 전략을 살펴봅니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: ai-agent-patterns</p><p><a href="https://archive.kreathlab.com/tech/ai-agent-patterns-03-tool-use">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 21 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai-agent</category>
      <category>llm</category>
      <category>architecture</category>
      <category>orchestration</category>
    </item>
    <item>
      <title><![CDATA[2장: AI 코드 리뷰 자동화 - 원리와 아키텍처]]></title>
      <link>https://archive.kreathlab.com/tech/ai-dev-workflow-02-code-review-principles</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-dev-workflow-02-code-review-principles</guid>
      <description><![CDATA[LLM이 코드를 이해하고 리뷰 피드백을 생성하는 원리를 분석하고, AI 코드 리뷰 시스템의 아키텍처를 설계합니다.]]></description>
      <content:encoded><![CDATA[<p>LLM이 코드를 이해하고 리뷰 피드백을 생성하는 원리를 분석하고, AI 코드 리뷰 시스템의 아키텍처를 설계합니다.</p><p><strong>AI / ML</strong> / 20분 읽기 / 시리즈: ai-dev-workflow</p><p><a href="https://archive.kreathlab.com/tech/ai-dev-workflow-02-code-review-principles">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 21 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>devtools</category>
      <category>automation</category>
      <category>code-quality</category>
      <category>devops</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[4장: LLM-as-Judge - LLM으로 LLM 평가하기]]></title>
      <link>https://archive.kreathlab.com/tech/llm-evaluation-04-llm-as-judge</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/llm-evaluation-04-llm-as-judge</guid>
      <description><![CDATA[LLM을 평가자로 활용하는 LLM-as-Judge 기법의 원리, 프롬프트 설계, 편향 완화 전략을 체계적으로 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>LLM을 평가자로 활용하는 LLM-as-Judge 기법의 원리, 프롬프트 설계, 편향 완화 전략을 체계적으로 다룹니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: llm-evaluation</p><p><a href="https://archive.kreathlab.com/tech/llm-evaluation-04-llm-as-judge">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 21 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>evaluation</category>
      <category>monitoring</category>
      <category>observability</category>
      <category>testing</category>
    </item>
    <item>
      <title><![CDATA[1장: MCP 소개와 핵심 개념]]></title>
      <link>https://archive.kreathlab.com/tech/mcp-guide-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/mcp-guide-01-introduction</guid>
      <description><![CDATA[Model Context Protocol이 무엇이고, 왜 AI 생태계의 표준으로 자리 잡았는지, 그리고 이 시리즈에서 다룰 내용의 전체 지도를 살펴봅니다.]]></description>
      <content:encoded><![CDATA[<p>Model Context Protocol이 무엇이고, 왜 AI 생태계의 표준으로 자리 잡았는지, 그리고 이 시리즈에서 다룰 내용의 전체 지도를 살펴봅니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: mcp-guide</p><p><a href="https://archive.kreathlab.com/tech/mcp-guide-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 21 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>mcp</category>
      <category>typescript</category>
      <category>python</category>
    </item>
    <item>
      <title><![CDATA[5장: 인덱싱과 검색 파이프라인 구축]]></title>
      <link>https://archive.kreathlab.com/tech/rag-system-05-indexing-retrieval</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/rag-system-05-indexing-retrieval</guid>
      <description><![CDATA[문서 로딩부터 임베딩 생성, 벡터 저장, 유사도 검색까지 RAG 파이프라인의 전체 흐름을 실제 코드로 구현합니다.]]></description>
      <content:encoded><![CDATA[<p>문서 로딩부터 임베딩 생성, 벡터 저장, 유사도 검색까지 RAG 파이프라인의 전체 흐름을 실제 코드로 구현합니다.</p><p><strong>AI / ML</strong> / 15분 읽기 / 시리즈: rag-system</p><p><a href="https://archive.kreathlab.com/tech/rag-system-05-indexing-retrieval">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 21 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>rag</category>
      <category>vector-database</category>
      <category>embedding</category>
      <category>retrieval</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[2장: 모델 서빙 프레임워크 - vLLM과 TGI 심층 비교]]></title>
      <link>https://archive.kreathlab.com/tech/ai-deployment-02-model-serving</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-deployment-02-model-serving</guid>
      <description><![CDATA[LLM 추론의 핵심 엔진인 vLLM과 Text Generation Inference를 아키텍처, 성능, 기능 측면에서 심층 비교하고 적합한 선택 기준을 제시합니다.]]></description>
      <content:encoded><![CDATA[<p>LLM 추론의 핵심 엔진인 vLLM과 Text Generation Inference를 아키텍처, 성능, 기능 측면에서 심층 비교하고 적합한 선택 기준을 제시합니다.</p><p><strong>AI / ML</strong> / 19분 읽기 / 시리즈: ai-deployment</p><p><a href="https://archive.kreathlab.com/tech/ai-deployment-02-model-serving">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 20 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>mlops</category>
      <category>kubernetes</category>
      <category>infrastructure</category>
      <category>performance</category>
    </item>
    <item>
      <title><![CDATA[4장: LoRA의 원리와 실전 적용]]></title>
      <link>https://archive.kreathlab.com/tech/fine-tuning-04-lora</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/fine-tuning-04-lora</guid>
      <description><![CDATA[LoRA(Low-Rank Adaptation)의 수학적 원리를 이해하고, 타겟 레이어 선택부터 하이퍼파라미터 튜닝까지 실전 적용법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>LoRA(Low-Rank Adaptation)의 수학적 원리를 이해하고, 타겟 레이어 선택부터 하이퍼파라미터 튜닝까지 실전 적용법을 다룹니다.</p><p><strong>AI / ML</strong> / 15분 읽기 / 시리즈: fine-tuning</p><p><a href="https://archive.kreathlab.com/tech/fine-tuning-04-lora">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 20 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>training</category>
      <category>mlops</category>
      <category>data-engineering</category>
    </item>
    <item>
      <title><![CDATA[1장: Next.js 15의 등장과 App Router의 진화]]></title>
      <link>https://archive.kreathlab.com/tech/nextjs-app-router-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/nextjs-app-router-01-introduction</guid>
      <description><![CDATA[Next.js 15부터 16까지의 진화를 조망합니다. App Router의 완성, 캐싱 전략 변화, Turbopack, Cache Components 등 핵심 변경사항을 개괄합니다.]]></description>
      <content:encoded><![CDATA[<p>Next.js 15부터 16까지의 진화를 조망합니다. App Router의 완성, 캐싱 전략 변화, Turbopack, Cache Components 등 핵심 변경사항을 개괄합니다.</p><p><strong>웹 개발</strong> / 10분 읽기 / 시리즈: nextjs-app-router</p><p><a href="https://archive.kreathlab.com/tech/nextjs-app-router-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 20 Jan 2026 00:00:00 GMT</pubDate>
      <category>web</category>
      <category>nextjs</category>
      <category>react</category>
      <category>typescript</category>
      <category>frontend</category>
    </item>
    <item>
      <title><![CDATA[5장: 구조화된 입력 - XML, JSON, 마크다운 활용]]></title>
      <link>https://archive.kreathlab.com/tech/prompt-engineering-05-structured-input</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/prompt-engineering-05-structured-input</guid>
      <description><![CDATA[프롬프트의 구조를 명확히 하는 XML, JSON, 마크다운 기반 입력 설계 기법과 모델별 최적 전략을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>프롬프트의 구조를 명확히 하는 XML, JSON, 마크다운 기반 입력 설계 기법과 모델별 최적 전략을 다룹니다.</p><p><strong>AI / ML</strong> / 18분 읽기 / 시리즈: prompt-engineering</p><p><a href="https://archive.kreathlab.com/tech/prompt-engineering-05-structured-input">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 20 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>prompt-engineering</category>
      <category>structured-output</category>
      <category>training</category>
    </item>
    <item>
      <title><![CDATA[3장: 구조적 패턴 매칭 실전 활용]]></title>
      <link>https://archive.kreathlab.com/tech/python-trends-03-pattern-matching</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/python-trends-03-pattern-matching</guid>
      <description><![CDATA[Python의 match/case 문을 실전에서 활용하는 방법을 다룹니다. 시퀀스, 매핑, 클래스 패턴부터 가드 조건, 중첩 패턴까지 실무 코드로 익힙니다.]]></description>
      <content:encoded><![CDATA[<p>Python의 match/case 문을 실전에서 활용하는 방법을 다룹니다. 시퀀스, 매핑, 클래스 패턴부터 가드 조건, 중첩 패턴까지 실무 코드로 익힙니다.</p><p><strong>프로그래밍</strong> / 14분 읽기 / 시리즈: python-trends</p><p><a href="https://archive.kreathlab.com/tech/python-trends-03-pattern-matching">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 20 Jan 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>python</category>
      <category>typescript</category>
      <category>performance</category>
      <category>devtools</category>
      <category>concurrency</category>
    </item>
    <item>
      <title><![CDATA[2장: ReAct 패턴 - 추론과 행동의 결합]]></title>
      <link>https://archive.kreathlab.com/tech/ai-agent-patterns-02-react-pattern</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-agent-patterns-02-react-pattern</guid>
      <description><![CDATA[ReAct 패턴의 원리와 구조를 이해하고, 추론-행동-관찰 루프를 직접 구현하여 LLM의 문제 해결 능력을 극대화하는 방법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>ReAct 패턴의 원리와 구조를 이해하고, 추론-행동-관찰 루프를 직접 구현하여 LLM의 문제 해결 능력을 극대화하는 방법을 다룹니다.</p><p><strong>AI / ML</strong> / 18분 읽기 / 시리즈: ai-agent-patterns</p><p><a href="https://archive.kreathlab.com/tech/ai-agent-patterns-02-react-pattern">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 19 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai-agent</category>
      <category>llm</category>
      <category>architecture</category>
      <category>orchestration</category>
    </item>
    <item>
      <title><![CDATA[1장: AI 기반 개발 워크플로우의 전체 그림]]></title>
      <link>https://archive.kreathlab.com/tech/ai-dev-workflow-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-dev-workflow-01-introduction</guid>
      <description><![CDATA[AI가 소프트웨어 개발의 각 단계를 어떻게 혁신하는지 전체 그림을 조망하고, 코드 리뷰, 테스트, 문서화, CI/CD 자동화의 핵심 개념을 정리합니다.]]></description>
      <content:encoded><![CDATA[<p>AI가 소프트웨어 개발의 각 단계를 어떻게 혁신하는지 전체 그림을 조망하고, 코드 리뷰, 테스트, 문서화, CI/CD 자동화의 핵심 개념을 정리합니다.</p><p><strong>AI / ML</strong> / 19분 읽기 / 시리즈: ai-dev-workflow</p><p><a href="https://archive.kreathlab.com/tech/ai-dev-workflow-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 19 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>devtools</category>
      <category>automation</category>
      <category>code-quality</category>
      <category>devops</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[3장: 자동 평가 파이프라인 구축]]></title>
      <link>https://archive.kreathlab.com/tech/llm-evaluation-03-automated-pipeline</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/llm-evaluation-03-automated-pipeline</guid>
      <description><![CDATA[코드 기반 메트릭과 벤치마크 자동화로 LLM 애플리케이션의 품질을 체계적으로 측정하는 평가 파이프라인을 구축합니다.]]></description>
      <content:encoded><![CDATA[<p>코드 기반 메트릭과 벤치마크 자동화로 LLM 애플리케이션의 품질을 체계적으로 측정하는 평가 파이프라인을 구축합니다.</p><p><strong>AI / ML</strong> / 20분 읽기 / 시리즈: llm-evaluation</p><p><a href="https://archive.kreathlab.com/tech/llm-evaluation-03-automated-pipeline">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 19 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>evaluation</category>
      <category>monitoring</category>
      <category>observability</category>
      <category>testing</category>
    </item>
    <item>
      <title><![CDATA[4장: 벡터 데이터베이스 비교와 선택]]></title>
      <link>https://archive.kreathlab.com/tech/rag-system-04-vector-databases</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/rag-system-04-vector-databases</guid>
      <description><![CDATA[Pinecone, Weaviate, Qdrant, pgvector 등 주요 벡터 데이터베이스의 특성을 비교하고 상황에 맞는 선택 가이드를 제공합니다.]]></description>
      <content:encoded><![CDATA[<p>Pinecone, Weaviate, Qdrant, pgvector 등 주요 벡터 데이터베이스의 특성을 비교하고 상황에 맞는 선택 가이드를 제공합니다.</p><p><strong>AI / ML</strong> / 15분 읽기 / 시리즈: rag-system</p><p><a href="https://archive.kreathlab.com/tech/rag-system-04-vector-databases">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 19 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>rag</category>
      <category>vector-database</category>
      <category>embedding</category>
      <category>retrieval</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[1장: AI 서비스 배포의 전체 그림과 핵심 과제]]></title>
      <link>https://archive.kreathlab.com/tech/ai-deployment-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-deployment-01-introduction</guid>
      <description><![CDATA[AI 모델을 프로덕션에 배포하기 위한 전체 파이프라인을 조망하고, 전통적 웹 서비스와 다른 AI 서비스만의 고유한 배포 과제를 분석합니다.]]></description>
      <content:encoded><![CDATA[<p>AI 모델을 프로덕션에 배포하기 위한 전체 파이프라인을 조망하고, 전통적 웹 서비스와 다른 AI 서비스만의 고유한 배포 과제를 분석합니다.</p><p><strong>AI / ML</strong> / 22분 읽기 / 시리즈: ai-deployment</p><p><a href="https://archive.kreathlab.com/tech/ai-deployment-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 18 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>mlops</category>
      <category>kubernetes</category>
      <category>infrastructure</category>
      <category>performance</category>
    </item>
    <item>
      <title><![CDATA[3장: 데이터 품질 관리와 전처리 파이프라인]]></title>
      <link>https://archive.kreathlab.com/tech/fine-tuning-03-data-preprocessing</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/fine-tuning-03-data-preprocessing</guid>
      <description><![CDATA[파인튜닝 학습 데이터의 정제, 중복 제거, 토큰화, 패딩 전략 등 실전 전처리 파이프라인을 구축하는 방법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>파인튜닝 학습 데이터의 정제, 중복 제거, 토큰화, 패딩 전략 등 실전 전처리 파이프라인을 구축하는 방법을 다룹니다.</p><p><strong>AI / ML</strong> / 18분 읽기 / 시리즈: fine-tuning</p><p><a href="https://archive.kreathlab.com/tech/fine-tuning-03-data-preprocessing">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 18 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>training</category>
      <category>mlops</category>
      <category>data-engineering</category>
    </item>
    <item>
      <title><![CDATA[4장: 역할 지정과 페르소나 설계]]></title>
      <link>https://archive.kreathlab.com/tech/prompt-engineering-04-role-persona</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/prompt-engineering-04-role-persona</guid>
      <description><![CDATA[LLM에게 전문가 역할을 부여하여 도메인 특화 응답을 이끌어내는 페르소나 설계의 원리와 실전 패턴을 체계적으로 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>LLM에게 전문가 역할을 부여하여 도메인 특화 응답을 이끌어내는 페르소나 설계의 원리와 실전 패턴을 체계적으로 다룹니다.</p><p><strong>AI / ML</strong> / 19분 읽기 / 시리즈: prompt-engineering</p><p><a href="https://archive.kreathlab.com/tech/prompt-engineering-04-role-persona">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 18 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>prompt-engineering</category>
      <category>structured-output</category>
      <category>training</category>
    </item>
    <item>
      <title><![CDATA[2장: 타입 파라미터 새 문법 - PEP 695]]></title>
      <link>https://archive.kreathlab.com/tech/python-trends-02-type-parameter-syntax</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/python-trends-02-type-parameter-syntax</guid>
      <description><![CDATA[Python 3.12에서 도입된 PEP 695 타입 파라미터 문법을 상세히 분석합니다. TypeVar의 간결한 선언, 제네릭 클래스와 함수의 새 문법, type 별칭을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>Python 3.12에서 도입된 PEP 695 타입 파라미터 문법을 상세히 분석합니다. TypeVar의 간결한 선언, 제네릭 클래스와 함수의 새 문법, type 별칭을 다룹니다.</p><p><strong>프로그래밍</strong> / 13분 읽기 / 시리즈: python-trends</p><p><a href="https://archive.kreathlab.com/tech/python-trends-02-type-parameter-syntax">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sun, 18 Jan 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>python</category>
      <category>typescript</category>
      <category>performance</category>
      <category>devtools</category>
      <category>concurrency</category>
    </item>
    <item>
      <title><![CDATA[1장: AI 에이전트의 등장과 설계 패턴 개요]]></title>
      <link>https://archive.kreathlab.com/tech/ai-agent-patterns-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/ai-agent-patterns-01-introduction</guid>
      <description><![CDATA[AI 에이전트가 무엇이고 왜 중요한지, 그리고 이 시리즈에서 다룰 핵심 설계 패턴들의 전체 지도를 살펴봅니다.]]></description>
      <content:encoded><![CDATA[<p>AI 에이전트가 무엇이고 왜 중요한지, 그리고 이 시리즈에서 다룰 핵심 설계 패턴들의 전체 지도를 살펴봅니다.</p><p><strong>AI / ML</strong> / 13분 읽기 / 시리즈: ai-agent-patterns</p><p><a href="https://archive.kreathlab.com/tech/ai-agent-patterns-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 17 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>ai-agent</category>
      <category>llm</category>
      <category>architecture</category>
      <category>orchestration</category>
    </item>
    <item>
      <title><![CDATA[2장: 평가 메트릭 설계 - 정확성, 관련성, 안전성]]></title>
      <link>https://archive.kreathlab.com/tech/llm-evaluation-02-metrics-design</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/llm-evaluation-02-metrics-design</guid>
      <description><![CDATA[LLM 애플리케이션의 품질을 수치화하는 핵심 메트릭을 설계하고, 작업 유형별로 적절한 메트릭을 선택하는 방법을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>LLM 애플리케이션의 품질을 수치화하는 핵심 메트릭을 설계하고, 작업 유형별로 적절한 메트릭을 선택하는 방법을 다룹니다.</p><p><strong>AI / ML</strong> / 19분 읽기 / 시리즈: llm-evaluation</p><p><a href="https://archive.kreathlab.com/tech/llm-evaluation-02-metrics-design">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 17 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>evaluation</category>
      <category>monitoring</category>
      <category>observability</category>
      <category>testing</category>
    </item>
    <item>
      <title><![CDATA[3장: 청킹 전략 - 문서 분할의 기술]]></title>
      <link>https://archive.kreathlab.com/tech/rag-system-03-chunking-strategies</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/rag-system-03-chunking-strategies</guid>
      <description><![CDATA[RAG 검색 품질을 좌우하는 청킹 전략의 종류, 벤치마크 결과, 그리고 최적의 청크 크기를 선택하는 실전 가이드입니다.]]></description>
      <content:encoded><![CDATA[<p>RAG 검색 품질을 좌우하는 청킹 전략의 종류, 벤치마크 결과, 그리고 최적의 청크 크기를 선택하는 실전 가이드입니다.</p><p><strong>AI / ML</strong> / 16분 읽기 / 시리즈: rag-system</p><p><a href="https://archive.kreathlab.com/tech/rag-system-03-chunking-strategies">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Sat, 17 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>rag</category>
      <category>vector-database</category>
      <category>embedding</category>
      <category>retrieval</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[2장: 학습 데이터 설계와 구축]]></title>
      <link>https://archive.kreathlab.com/tech/fine-tuning-02-data-design</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/fine-tuning-02-data-design</guid>
      <description><![CDATA[파인튜닝 성패를 좌우하는 학습 데이터의 설계 원칙, 수집 전략, 데이터 형식을 실전 관점에서 체계적으로 안내합니다.]]></description>
      <content:encoded><![CDATA[<p>파인튜닝 성패를 좌우하는 학습 데이터의 설계 원칙, 수집 전략, 데이터 형식을 실전 관점에서 체계적으로 안내합니다.</p><p><strong>AI / ML</strong> / 20분 읽기 / 시리즈: fine-tuning</p><p><a href="https://archive.kreathlab.com/tech/fine-tuning-02-data-design">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 16 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>training</category>
      <category>mlops</category>
      <category>data-engineering</category>
    </item>
    <item>
      <title><![CDATA[3장: Chain-of-Thought 추론 기법]]></title>
      <link>https://archive.kreathlab.com/tech/prompt-engineering-03-chain-of-thought</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/prompt-engineering-03-chain-of-thought</guid>
      <description><![CDATA[LLM에게 단계적 사고를 유도하는 Chain-of-Thought 프롬프팅의 원리, 변형 기법, 그리고 최신 추론 모델에서의 활용 전략을 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>LLM에게 단계적 사고를 유도하는 Chain-of-Thought 프롬프팅의 원리, 변형 기법, 그리고 최신 추론 모델에서의 활용 전략을 다룹니다.</p><p><strong>AI / ML</strong> / 18분 읽기 / 시리즈: prompt-engineering</p><p><a href="https://archive.kreathlab.com/tech/prompt-engineering-03-chain-of-thought">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 16 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>prompt-engineering</category>
      <category>structured-output</category>
      <category>training</category>
    </item>
    <item>
      <title><![CDATA[1장: Python 3.12~3.13, 무엇이 달라졌는가]]></title>
      <link>https://archive.kreathlab.com/tech/python-trends-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/python-trends-01-introduction</guid>
      <description><![CDATA[Python 3.12와 3.13에서 도입된 핵심 변화를 조망하고, 타입 시스템 개선부터 GIL 제거, 차세대 도구 생태계까지 시리즈의 전체 로드맵을 제시합니다.]]></description>
      <content:encoded><![CDATA[<p>Python 3.12와 3.13에서 도입된 핵심 변화를 조망하고, 타입 시스템 개선부터 GIL 제거, 차세대 도구 생태계까지 시리즈의 전체 로드맵을 제시합니다.</p><p><strong>프로그래밍</strong> / 11분 읽기 / 시리즈: python-trends</p><p><a href="https://archive.kreathlab.com/tech/python-trends-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Fri, 16 Jan 2026 00:00:00 GMT</pubDate>
      <category>programming</category>
      <category>python</category>
      <category>typescript</category>
      <category>performance</category>
      <category>devtools</category>
      <category>concurrency</category>
    </item>
    <item>
      <title><![CDATA[1장: LLM 평가의 필요성과 전체 프레임워크]]></title>
      <link>https://archive.kreathlab.com/tech/llm-evaluation-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/llm-evaluation-01-introduction</guid>
      <description><![CDATA[LLM 애플리케이션을 프로덕션에서 안정적으로 운영하기 위해 평가와 모니터링이 왜 필수인지, 전체 프레임워크를 체계적으로 소개합니다.]]></description>
      <content:encoded><![CDATA[<p>LLM 애플리케이션을 프로덕션에서 안정적으로 운영하기 위해 평가와 모니터링이 왜 필수인지, 전체 프레임워크를 체계적으로 소개합니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: llm-evaluation</p><p><a href="https://archive.kreathlab.com/tech/llm-evaluation-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 15 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>evaluation</category>
      <category>monitoring</category>
      <category>observability</category>
      <category>testing</category>
    </item>
    <item>
      <title><![CDATA[2장: 임베딩 모델의 이해와 선택]]></title>
      <link>https://archive.kreathlab.com/tech/rag-system-02-embedding-models</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/rag-system-02-embedding-models</guid>
      <description><![CDATA[텍스트 임베딩의 원리부터 2026년 최신 모델 벤치마크, 프로덕션 환경에서의 선택 기준까지 체계적으로 안내합니다.]]></description>
      <content:encoded><![CDATA[<p>텍스트 임베딩의 원리부터 2026년 최신 모델 벤치마크, 프로덕션 환경에서의 선택 기준까지 체계적으로 안내합니다.</p><p><strong>AI / ML</strong> / 17분 읽기 / 시리즈: rag-system</p><p><a href="https://archive.kreathlab.com/tech/rag-system-02-embedding-models">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Thu, 15 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>rag</category>
      <category>vector-database</category>
      <category>embedding</category>
      <category>retrieval</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[1장: 파인튜닝의 개념과 사전 학습 모델의 이해]]></title>
      <link>https://archive.kreathlab.com/tech/fine-tuning-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/fine-tuning-01-introduction</guid>
      <description><![CDATA[LLM 파인튜닝이 무엇인지, 사전 학습 모델과 어떤 관계가 있는지, 언제 파인튜닝이 필요한지를 체계적으로 정리합니다.]]></description>
      <content:encoded><![CDATA[<p>LLM 파인튜닝이 무엇인지, 사전 학습 모델과 어떤 관계가 있는지, 언제 파인튜닝이 필요한지를 체계적으로 정리합니다.</p><p><strong>AI / ML</strong> / 21분 읽기 / 시리즈: fine-tuning</p><p><a href="https://archive.kreathlab.com/tech/fine-tuning-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 14 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>training</category>
      <category>mlops</category>
      <category>data-engineering</category>
    </item>
    <item>
      <title><![CDATA[2장: 제로샷과 퓨샷 프롬프팅]]></title>
      <link>https://archive.kreathlab.com/tech/prompt-engineering-02-zero-few-shot</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/prompt-engineering-02-zero-few-shot</guid>
      <description><![CDATA[예시 없이 지시하는 제로샷부터 예시를 활용하는 퓨샷까지, 예시 기반 프롬프팅의 원리와 실전 활용 전략을 체계적으로 다룹니다.]]></description>
      <content:encoded><![CDATA[<p>예시 없이 지시하는 제로샷부터 예시를 활용하는 퓨샷까지, 예시 기반 프롬프팅의 원리와 실전 활용 전략을 체계적으로 다룹니다.</p><p><strong>AI / ML</strong> / 21분 읽기 / 시리즈: prompt-engineering</p><p><a href="https://archive.kreathlab.com/tech/prompt-engineering-02-zero-few-shot">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Wed, 14 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>prompt-engineering</category>
      <category>structured-output</category>
      <category>training</category>
    </item>
    <item>
      <title><![CDATA[1장: RAG 시스템 소개와 아키텍처 개요]]></title>
      <link>https://archive.kreathlab.com/tech/rag-system-01-introduction</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/rag-system-01-introduction</guid>
      <description><![CDATA[검색 증강 생성(RAG)이 무엇이고 왜 필요한지, 핵심 구성 요소와 전체 아키텍처를 체계적으로 살펴봅니다.]]></description>
      <content:encoded><![CDATA[<p>검색 증강 생성(RAG)이 무엇이고 왜 필요한지, 핵심 구성 요소와 전체 아키텍처를 체계적으로 살펴봅니다.</p><p><strong>AI / ML</strong> / 14분 읽기 / 시리즈: rag-system</p><p><a href="https://archive.kreathlab.com/tech/rag-system-01-introduction">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Tue, 13 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>rag</category>
      <category>vector-database</category>
      <category>embedding</category>
      <category>retrieval</category>
      <category>llm</category>
    </item>
    <item>
      <title><![CDATA[1장: 프롬프트 엔지니어링의 기초와 핵심 원칙]]></title>
      <link>https://archive.kreathlab.com/tech/prompt-engineering-01-fundamentals</link>
      <guid isPermaLink="true">https://archive.kreathlab.com/tech/prompt-engineering-01-fundamentals</guid>
      <description><![CDATA[프롬프트 엔지니어링이 무엇이고 왜 중요한지, LLM과의 효과적인 소통을 위한 핵심 원칙과 기초 개념을 체계적으로 살펴봅니다.]]></description>
      <content:encoded><![CDATA[<p>프롬프트 엔지니어링이 무엇이고 왜 중요한지, LLM과의 효과적인 소통을 위한 핵심 원칙과 기초 개념을 체계적으로 살펴봅니다.</p><p><strong>AI / ML</strong> / 20분 읽기 / 시리즈: prompt-engineering</p><p><a href="https://archive.kreathlab.com/tech/prompt-engineering-01-fundamentals">전체 글 읽기</a></p>]]></content:encoded>
      <author>hello@kreathlab.com (Kreath)</author>
      <pubDate>Mon, 12 Jan 2026 00:00:00 GMT</pubDate>
      <category>ai</category>
      <category>llm</category>
      <category>prompt-engineering</category>
      <category>structured-output</category>
      <category>training</category>
    </item>
  </channel>
</rss>