Evaluating AI applications, not just models.
A London meetup and community for people who build AI-powered products. Model benchmarks on leaderboards are interesting, but how do you evaluate your specific RAG pipeline in production?
Evaluating retrieval quality and end-to-end RAG pipelines in production
How to catch and fix problems in deployed AI systems
Evaluation approaches that combine human expertise with automation
Integrating evals into development and CI/CD pipelines
You deploy and monitor models in production
You integrate LLMs or RAG into an existing product
You measure and improve the performance of AI systems
You need to understand the quality of AI features
You study the reliability and safety of AI
Practical interest and curiosity are all you need
"What broke in production and how we fixed it"— no sales pitches