[officehours 56] Grading AI with Evals

​Join us for [officehours 56] Grading AI with Evals, a practical conversation on how to measure and improve AI systems when building agents and harnesses is more accessible than ever. We will host a few speakers who will share what they have learned from using evals in real-world workflows, with a focus on how teams and individual builders can apply these ideas in a fast-moving environment.

​We will dig into three core questions: what should actually be measured, how to choose an effective eval strategy, and how to turn evaluation into a continuous improvement loop. If you are working on AI products, experimenting with agents, or trying to make model behavior more reliable, this session will offer clear perspectives on how to evaluate progress in a way that leads to better decisions and better systems.

​More to come.