Introducing Arize-Phoenix and OpenInference for AI Evaluation and Observability

Introducing Arize-Phoenix and OpenInference for AI Evaluation and Observability

Explore the critical need for observability and evaluation in generative AI applications through this 43-minute conference talk by Mikyo King, Head of Open Source & Founding Engineer at Arize AI. Discover OpenInference, a new open set of plugins and conventions for tracing AI applications, and Arize-Phoenix, a platform offering AI evaluation and observability. Gain practical insights through real-world examples and use cases, including how to incorporate LLMs as judges to guard against hallucinations and poor retrieval. Learn from King's expertise as a former Apple engineer and understand the challenges of working with probabilistic LLMs in deterministic environments.