Evaluating AI Agents скачать в хорошем качестве

Evaluating AI Agents

AI agents have taken over the ecosystem during 2024, but very few real agents make it to production. Making an MVP agent is easier than ever, but evaluating and improving them is the true challenge. In this talk, we'll dive into techniques for evaluating agents, which evaluators to use on which pieces of your pipeline, and how you should prepare your agent for production. About the Speaker: John Gilhuly, a leading voice in AI Observability, is the Developer Advocate at Arize AI. With extensive expertise in open-source AI, technical partnerships, and community building, John focuses on driving effective evaluation strategies for AI agents and advancing observability practices across the industry. Join this session to gain insights from a seasoned advocate dedicated to enhancing AI agent performance and reliability in real-world applications.