RAG Pipeline
Retrieval-Augmented Generation combines everything in this section into one flow: load documents, split them, embed them into a searchable index, then at query time retrieve the relevant chunks and hand them to a model alongside the question.