The fastest way to understand RAG is to build the simple version and break it. You will wire up the basic loop: embed your documents, store the vectors, embed the user's question, retrieve the closest chunks, and stuff them into the prompt.

It works, then it fails in instructive ways: it retrieves irrelevant chunks, misses the obvious answer, or returns something confidently wrong. Those failures are the point. Every technique in the rest of the session exists to fix a specific way naive RAG breaks.