Finch

Finch is a memory solution for my AI agent system. It achieves SOTA performance on the MEME benchmark with an overall score of 65.3% and scores 71.3% on the Enterprise-RAG-Bench.

Enterprise-RAG-BenchMEME

Finch uses the judge and reader each benchmark specifies. I cannot speak for the others.

Currently in the process of writing a detailed breakdown which will be called Memory for AI where I conduct a mini case study covering current memory systems, inspiration for the project, how I chose to measure success/performance, latency optimizations, effects of the LLM you choose to use in this memory system, and more.