Giuliano Morse

UX Researcher

Meta

Giuliano Morse is a Staff UX Researcher at Meta Superintelligence Labs, where he develops new ways for UX research to directly shape frontier AI models. Over seven years at Meta, he has led three UX research teams within FAIR, helped shape AI products used by millions, and developed research methods that have fundamentally changed how AI models are evaluated and trained. Most notably, he developed User Evals, a methodology for evaluating model quality with real end users, that became a core signal for Muse and Llama model post-training. His work has driven significant improvements in user engagement and helped scale user-centered evaluation across Meta’s AI organization. Giuliano is passionate about expanding the influence of UXR from informing product decisions to directly shaping the intelligence behind AI systems.

Sessions

Oct 27

4:05 PM - 4:25 PM PDT

Green Stage

Customer Experience & Journey

Media

Entertainment & Telecom

Human Evals in the AI Age

AI development benefits from real users' feedback captured via natural experiences.  Human Evals are delivering quantitative and qualitative rigor, at the speed of product, with real people. This continuous learning from the users’ perspective is now integral to Meta's Superintelligence Labs. Join us to learn how it works and insights gained.

Key Takeaways:

1. Human Evals provide a complete, honest perspective of the AI UX at a personal level.

2. Direct competitive comparison uncovers unforeseen strengths & gaps via custom, concrete benchmarks.

3. People’s expectations of AI are evolving at the pace of development; Evals catch early shifts.

Also speaking

Lisa Phelan

Cofounder, F'inn

Speaker