Retrieval-Augmented Generation in Practice
Build a RAG system that actually retrieves the right thing: chunking that preserves meaning, hybrid search, re-ranking, and answers you can trace to a source.
Advanced2h 17m4 modules7.4k students
What you'll learn
Design a retrieval pipeline
Understand every stage from ingestion to grounded answer, and where each one typically fails.
Chunk and embed sensibly
Split documents so a retrieved passage still makes sense on its own, and pick an embedding model deliberately.
Measure retrieval quality
Score recall and precision so you can tell whether a change to the pipeline helped.
Ground every answer
Cite sources, detect unsupported claims, and make the system say it does not know.
Course Content
4 modules • 2h 17m
Your Progress
Not started