Builds retrieval-augmented generation pipelines end to end: document ingestion, structure-aware chunking, embeddings, vector storage, hybrid retrieval, and grounded answer generation with citations.
