← BACK TO KNOWLEDGE

KNOWLEDGE / RAG / Retrieval

RAG Retrieval and Reranking

How candidate generation and reranking work together to produce a small, relevant context window.

RAGRetrievalReranking
DOMAIN
AI Engineering
LEVEL
Advanced
READ
9 min
UPDATED
Aug 27, 2026

MENTAL MODEL / KEY IDEAS

Keep these in mind

  1. 01Optimize recall before precision
  2. 02Reranking is a separate relevance model
  3. 03Evaluate with real queries
01

Two different jobs

Retrieval quickly finds a broad candidate set. Reranking spends more compute to order those candidates by relevance to the exact question.

  • Fast first stage
  • Precise second stage
  • Small final context
02

Hybrid retrieval

Lexical search preserves exact terms while embeddings recover semantic matches. Combining both is often more robust than treating either as universal.

  • Normalize scores
  • Deduplicate passages
  • Keep source metadata
03

Evaluation

Measure retrieval before answer quality. If the supporting passage never reaches the model, generation tuning cannot repair the system.