RAG systems tutorial roadmap

A tutorial outline for building retrieval-augmented generation systems with evaluation from the start.

Tutorial Goal

Build a retrieval-augmented generation workflow that is evaluated like a system, not demoed like a prompt.

Sections To Build

1. Document ingestion and cleaning.
2. Chunking and metadata design.
3. Embedding selection.
4. FAISS or Pinecone indexing.
5. Retrieval, reranking, and response synthesis.
6. Citation and groundedness checks.
7. Regression tests for prompt and retrieval changes.

Notebook Plan

• notebooks/rag-systems/01-ingestion.ipynb
• notebooks/rag-systems/02-indexing.ipynb
• notebooks/rag-systems/03-generation.ipynb
• notebooks/rag-systems/04-evaluation.ipynb