RAG

The path from raw documents to useful model context: embeddings, vector search, retrieval quality, reranking, and answer generation.