conv.

All stories
AIQuiet 13d · day 19

Chandula Senevirathna releases agentic RAG tutorial with LangGraph

Ten Jupyter notebooks teach retrieval-augmented generation patterns that let AI agents correct their own mistakes and defer to humans.

newest coverage 13d agosummary written 13d ago12 pieces59 likes & upvotes35 comments Hacker NewsMastodon

What to know

  • A tutorial repository with ten Jupyter notebooks on agentic RAG using LangGraph teaches how to build AI agents that can self-correct, request human input, parallelize work, and plan ahead—addressing core limitations of traditional retrieval-augmented generation.
  • Four free notebooks cover foundational LangGraph concepts (state, reducers, conditional routing) and a working agentic RAG agent; six paid notebooks on Gumroad explore advanced patterns on a spectrum of autonomy and verification.
  • The material has been reposted to Hacker News at least eight times since early September under varying titles emphasizing different angles: practical patterns, limitation of traditional RAG, hallucination risks, and reliability concerns.

Chandula Senevirathna Developer and course creator

Chandula Senevirathna releases agentic RAG tutorial with LangGraph
github.com

How it unfolded 10 developments, newest first · click a bar or a number to jump posts

Peak 3 pieces in 5h at Sep 14, 10 AM; 12 pieces over 20 days (12 posts) Sep 8, 2 PM — 1 piece · 1 post — Hacker News 1Sep 8, 7 PM — quietSep 9, 12 AM — 1 piece · 1 post — Hacker News 1Sep 9, 5 AM — quietSep 9, 10 AM — 1 piece · 1 post — Hacker News 1Sep 9, 3 PM — quietSep 9, 8 PM — quietSep 10, 1 AM — 1 piece · 1 post — Hacker News 1Sep 10, 6 AM — quietSep 10, 11 AM — quietSep 10, 4 PM — 1 piece · 1 post — Hacker News 1Sep 10, 9 PM — quietSep 11, 2 AM — quietSep 11, 7 AM — quietSep 11, 12 PM — quietSep 11, 5 PM — quietSep 11, 10 PM — quietSep 12, 3 AM — quietSep 12, 8 AM — quietSep 12, 1 PM — quietSep 12, 6 PM — 1 piece · 1 post — Hacker News 1Sep 12, 11 PM — quietSep 13, 4 AM — 1 piece · 1 post — Hacker News 1Sep 13, 9 AM — 1 piece · 1 post — Hacker News 1Sep 13, 2 PM — quietSep 13, 7 PM — 1 piece · 1 post — Hacker News 1Sep 14, 12 AM — quietSep 14, 5 AM — quietSep 14, 10 AM — 3 pieces · 3 posts — Hacker News 2, Mastodon 1Sep 14, 3 PM — quietSep 14, 8 PM — quietSep 15, 1 AM — quietSep 15, 6 AM — quietSep 15, 11 AM — quietSep 15, 4 PM — quietSep 15, 9 PM — quietSep 16, 2 AM — quietSep 16, 7 AM — quietSep 16, 12 PM — quietSep 16, 5 PM — quietSep 16, 10 PM — quietSep 17, 3 AM — quietSep 17, 8 AM — quietSep 17, 1 PM — quietSep 17, 6 PM — quietSep 17, 11 PM — quietSep 18, 4 AM — quietSep 18, 9 AM — quietSep 18, 2 PM — quietSep 18, 7 PM — quietSep 19, 12 AM — quietSep 19, 5 AM — quietSep 19, 10 AM — quietSep 19, 3 PM — quietSep 19, 8 PM — quietSep 20, 1 AM — quietSep 20, 6 AM — quietSep 20, 11 AM — quietSep 20, 4 PM — quietSep 20, 9 PM — quietSep 21, 2 AM — quietSep 21, 7 AM — quietSep 21, 12 PM — quietSep 21, 5 PM — quietSep 21, 10 PM — quietSep 22, 3 AM — quietSep 22, 8 AM — quietSep 22, 1 PM — quietSep 22, 6 PM — quietSep 22, 11 PM — quietSep 23, 4 AM — quietSep 23, 9 AM — quietSep 23, 2 PM — quietSep 23, 7 PM — quietSep 24, 12 AM — quietSep 24, 5 AM — quietSep 24, 10 AM — quietSep 24, 3 PM — quietSep 24, 8 PM — quietSep 25, 1 AM — quietSep 25, 6 AM — quietSep 25, 11 AM — quietSep 25, 4 PM — quietSep 25, 9 PM — quietSep 26, 2 AM — quietSep 26, 7 AM — quietSep 26, 12 PM — quietSep 26, 5 PM — quietSep 26, 10 PM — quietYesterday, 3 AM — quietYesterday, 8 AM — quietYesterday, 1 PM — quietYesterday, 6 PM — quietYesterday, 11 PM — quiet 1–2345678–10
Sep 9Sep 11Sep 13Sep 15Sep 17Sep 19Sep 21Sep 23Sep 25now · 2:30 AM ET
  1. 10

    Gumroad course reposted with emphasis on RAG reliability

    The Gumroad paid course is reposted on Hacker News under the title "Is Your RAG Confidently Wrong?," scoring 6 points with 5 comments, framing the material as addressing overconfident but incorrect outputs from RAG systems.

  2. 9

    Content shared to Mastodon

    The repository is shared to Mastodon with the title "RAG Might Be Lying to You," extending distribution beyond Hacker News.

    • hkrn@mstdn.social

      RAG Might Be Lying to You L: https:// github.com/ChandulaSenevirathn a/Agentic_RAG C: https:// news.ycombinator.com/item?id=4 9698744 posted on 2026.09.14 at 11:35:25 (c=1, p=6)

      hkrn@mstdn.socialMastodon13d agoview on Mastodon ↗
  3. 8

    Repository reposted with cautionary framing

    The GitHub repository appears again on Hacker News with the title "RAG Might Be Lying to You," scoring 7 points with 2 comments, emphasizing the problem of hallucination and unreliable retrieval in standard RAG systems.

  4. 7

    Repository reposted emphasizing traditional RAG limitations

    The GitHub repository is reposted on Hacker News with the title "Why Beyond Traditional RAG?," reaching 5 points with 1 comment.

  5. 6

    Another repost of Gumroad course with alternative title

    The paid course is reposted on Hacker News with a slightly different title, "Why Beyond the Waters of Traditional RAG," scoring 4 points.

  6. 5

    Gumroad page reposted with consistent engagement

    The Gumroad course link is reposted on Hacker News under the same title "Why Beyond Traditional RAG?," again reaching 8 points with 4 comments.

  7. 1 day quiet
  8. 4

    Gumroad course page posted to Hacker News

    A Hacker News post linking to the paid course on Gumroad ("Why Beyond Traditional RAG?") appears, scoring 6 points with 7 comments, indicating community interest in the paid material.

  9. 3

    Post resurfaces on Hacker News with refined framing

    The repository is reposted on Hacker News under the title "Practical Agentic RAG patterns implemented with LangGraph," gaining 8 points and 7 comments. The post description highlights the notebook progression and the spectrum of patterns covered by the paid notebooks.

    “Ten self-contained Jupyter notebooks. The first four are free and build up the core LangGraph primitives and a working agentic RAG agent from scratch; the remaining six are a paid, deeper dive into more advanced ways of making a Retrieval-Augmented Generation (RAG) pipeline "agentic" — able to correct its own mistakes, defer to a person, split work across specialists, run branches in parallel, or plan ahead.”
    — Repository documentation · source
  10. 2

    Repository gains initial traction on Hacker News

    The repository is submitted to Hacker News under the title "Advanced RAG," reaching a score of 4 with early engagement as developers discover the tutorial.

  11. 1

    Senevirathna publishes agentic RAG repository on GitHub

    A GitHub repository containing ten Jupyter notebooks on practical agentic RAG patterns implemented with LangGraph is made public. The first four notebooks are free and cover core LangGraph primitives (state, reducers, conditional routing) leading to a working agentic RAG agent; six additional paid notebooks explore advanced patterns.