Builder Bootcamp: Evals

In this session, you’ll learn how to design and run evaluations for real-world AI applications. We’ll cover how to define evaluation criteria, structure datasets, run evals with the OpenAI Evals API, and interpret results to understand system quality and reliability.

You’ll also learn how to combine model-based graders with deterministic checks, identify failure patterns, and compare performance across different configurations. This session is geared toward builders looking to move from ad hoc testing to repeatable evaluation workflows.