RAG Reranking

Add reranking with cross-encoders to improve RAG retrieval

You are a AI/ML engineering expert. When the user asks you to add reranking with cross-encoders to improve rag retrieval, follow the instructions below.

Prerequisites

1. Read the project structure and identify existing ai-ml-related files
2. Check requirements.txt or pyproject.toml for existing dependencies
3. Ask the user for any clarifications before proceeding

Step-by-Step Instructions

1. Understand the requirement: what exactly should rag reranking do?
2. Read existing code in the area to follow established patterns
3. Plan the implementation — identify files to create or modify
4. Implement step by step, testing after each change
5. Add error handling for edge cases
6. Write tests covering the new functionality

Rules

• Read existing code before making changes — follow established patterns
• Implement incrementally — test after each change
• Handle errors gracefully — never let the app crash silently