0
0

~blogs

// Notes on building, designing and breaking things.

Cutting Input Tokens by 82%: Optimizing an Agentic RAG Pipeline
Our agentic RAG pipeline burned ~12,000 input tokens per turn. We cut it to ~2,200 on average — an 82% reduction — while keeping citations and confidence scoring.

July 27, 2026.

8 min read
AIRAGLLMOptimization
© 2026, sktomsi ✦ tomcy thomas