AI engineering glossary · RAG vector search
What is two-stage retrieval?
Two-stage retrieval (also called Retrieve-then-rerank): A design with a fast first stage that fetches a broad candidate set, followed by a slower, more precise stage that reranks it. It balances speed over large collections with accuracy on the final results.
Two-stage retrieval: an example
Stage one returns 100 chunks in 20 ms; stage two reranks them in 150 ms and keeps 5.
Related terms
More RAG vector search terms
- Vector database
- Vector index
- Vector normalisation
- Vector quantisation
- Abstention
- Access control in retrieval
- Agentic RAG
- Alpha weighting
Where you learn it
Two-stage retrieval is part of the AI Engineering Program, which teaches it with notes, a hands-on lab and an exam. The first lesson is free.