Retrieval Augmented Generation (RAG)

Language models answer questions about documents they have never seen when a retrieval step gives them the right passages. This path covers the full RAG pipeline, from chunking and vector search to hybrid retrieval, grounded answers and evaluation. It then adds knowledge graphs with GraphRAG, for questions that connect several facts or cover a whole collection of documents.