Asked about The RAG pipeline

How would you chunk documents?

An answer that works

On structure rather than a fixed character count: paragraphs and sections, keeping headings attached to their content, with slight overlap between chunks. Store the document title and heading path with each chunk and include them when embedding, so a chunk is self-describing.

Reading an answer is not being able to give one. The section that teaches this is The RAG pipeline, 4 hours of Retrieval-augmented generation.

Learn this properly

Also asked about Retrieval-augmented generation

What problem does RAG solve?Why ask the model to cite sources?When would you fine-tune instead of using RAG?Your RAG system gives a wrong answer. How do you find out why?How do you handle a question your documents cannot answer?What is query rewriting and why does it matter?How do you keep a RAG index current?