Sign in

Week 3LLM Core

5 lessons · project on day 18

Week progress0 / 5
  1. Day 13

    Prompt Engineering

    Prompting as engineering, not folklore — structured output, schema-constrained generation, and evals that survive a model upgrade.

    Open
  2. Day 14

    Embeddings

    What a vector actually is, cosine similarity, and why chunking strategy — not the embedding model — is where most RAG systems fail.

    Open
  3. Day 15

    Vector Database

  4. Day 16

    RAG

  5. Day 17

    MCP

Day 18 project

Docs RAG Assistant

Point it at a documentation site, chunk and embed the pages, and answer questions with inline citations back to the source section. The interesting work is not the retrieval call — it is the chunking strategy, and discovering how badly naive fixed-size chunks perform against headings-aware ones.

Stack

  • TypeScript
  • pgvector
  • OpenAI-compatible embeddings
  • Next.js
  • Zod

The hard parts

  • Chunk boundaries that split a code sample or a table in half, which quietly destroys answer quality without ever throwing an error.
  • Deciding when retrieval found nothing useful — a cosine similarity of 0.62 looks fine numerically and is often garbage.
  • Citations that survive re-indexing, which means chunk IDs must be stable across runs rather than array positions.

Why it belongs in your portfolio

Everyone claims RAG experience. A repo with a chunking-strategy comparison and a retrieval evaluation set shows you have actually measured it instead of shipping the first thing that returned text.