rag-optimization

What problem does it solve?

RAG pipelines can be brittle; this Skill helps adjust search weighting, reranking, query expansion, and cache strategies to improve relevance and reduce latency.

Core Features & Use Cases

• Search optimization: balance vector/keyword/graph weights.
• Re-ranking: adjust Cross-Encoder thresholds; Top-K results.
• Query expansion: Multi-query expansion counts; variation quality.
• Caching: assess semantic cache hit rates and invalidate when needed.

Quick Start

Tune the retrieval weights (set vector_weight=0.6, keyword_weight=0.4), enable query expansion (count=5), then run a sample RAG query to compare results.