← Learn
Learn · AI

Retrieval and RAG

Seven lessons on retrieval-augmented generation: how embeddings and cosine similarity find passages by meaning, how to chunk documents, vector stores, hybrid keyword and vector search with reranking, answering with citations, measuring it all, and the common ways it fails, including instructions hidden inside documents.

  1. Embeddings and similarity

    Turning text into vectors so that similar meanings end up close together, and measuring closeness with cosine similarity.

    6 questions

  2. Chunking documents

    Splitting documents into passages that can be found and understood on their own: size, overlap, structure and metadata.

    6 questions

  3. Vector stores and search

    Where the vectors live, exact and approximate nearest-neighbour search, top-k, filters and keeping the index up to date.

    6 questions

  4. Hybrid search and reranking

    Why keyword search still matters, combining it with vector search, and reranking the best candidates.

    6 questions

  5. Answering with sources

    Putting retrieved chunks in the prompt, asking for citations, saying "I don't know", and checking the citations in code.

    6 questions

  6. Evaluating RAG

    Measuring retrieval and answers separately: recall@k, MRR, faithfulness and correctness.

    6 questions

  7. When RAG goes wrong

    Bad chunks, a stale index, missed matches, hallucination, and instructions hidden inside documents.

    6 questions
More info

How it works

Each lesson is a short read with examples, then a few practice questions. Some ask you to type an answer (a number, an address, what a piece of code prints), some have you pick one, and some have you tap steps or lines into the right order. You can try again as often as you like; after two misses there’s a button to show the answer.

What do I need to know first?

How LLMs work first, or knowing what tokens, prompts and context windows are. The examples are short Python you only need to read.

The other AI and data engineering courses

Where is my progress saved? In this browser, like the rest of the site. Nothing to sign up for.

“Python” is a registered trademark of the Python Software Foundation. This course isn’t made by or affiliated with it.