Measure what matters: Intro to AI evals for common use cases

​This session will break down the basics through practical examples that are suitable for both AI engineers and PMs.

​We'll work through how to evaluate three common use cases:

• ​Customer support agent: "Is my AI support agent ready for customers?"

• ​Content/code generation: "How do I know if my AI's output is actually good?"

• ​Prompt optimization and model testing: "Which version of my AI setup works better?"

​We'll also cover how to write good scoring functions and manage datasets. No prior evaluation experience is required. Framework and model-agnostic approaches that work with any AI application.