Build LLM Apps with RAG

Connect a language model to your own documents with retrieval, and measure answer quality.

Build a retrieval-augmented generation app from scratch: chunking, embeddings, vector search, and generation with Gemini. Then evaluate answers so you know when the app is ready.

What you will be able to do

• Build a full RAG pipeline over your own documents
• Choose chunking and retrieval settings with data
• Evaluate answer quality and spot hallucinations

Before you start

• Python basics
• A free Gemini API key

Course outline

1. Chunking and embeddings: Split documents and turn them into vectors. A searchable index of your documents.
2. Retrieval that finds the right passage: Vector search with ranking checks. Retrieval metrics per question.
3. Answers with citations: Generation with Gemini that cites its sources. A working RAG app.
4. Know when it's ready: An answer quality eval set. A release check you run before each change.

All AI and machine learning courses