Module 1: RAG Architecture & Document Chunking

Retrieval-Augmented Generation combines search with LLMs for grounded answers.

What is RAG?

RAG retrieves relevant context before generating, reducing hallucinations.

Chunking Strategies

Compare fixed-size, semantic, and recursive chunking for different document types.

Metadata Enrichment

Add source, page number, and section headers to improve retrieval accuracy.

Evaluation Metrics

Measure chunk quality with recall@k before building the full pipeline.