Data: 19 maio
Hora: 13h
Local: Anfiteatro PA2, Pavilhão de Matemática, Campus Alameda
Orador: Pedro Saleiro (Opnova)
Título: “Computer Use: The Last Frontier of LLM Agents”
Resumo:
LLMs now ace coding tasks or olympiad math, but still can’t reliably finish a routine back-office task, let alone repeat it thousands of times. I believe computer use is the last frontier of LLM agents, and probably the most consequential one in regulated industries. In this talk I’ll cover the state of the art in computer-use agents, from visual grounding and GUI world models to reinforcement learning on agentic environments. I’ll also get into the alignment risks that show up when agents stop suggesting and start acting. Finally, I’ll share what we’ve learned at Opnova deploying computer-use agents for enterprise operations in regulated banking, where reliability, auditability, and continual learning aren’t optional.
Nota biográfica:
Pedro Saleiro is Co-Founder and Chief AI Officer at Opnova, a startup building AI computer-use agents for enterprise operations in regulated industries. Before Opnova, Pedro led the FATE and AI Research groups at Feedzai, where his teams filed numerous patents, published at top AI venues, and shipped multiple research innovations to product. His career includes a postdoctoral position at the University of Chicago with Rayid Ghani, a Ph.D. internship at Microsoft Research, and a Ph.D. in Informatics Engineering from the Faculty of Engineering at the University of Porto. He has created several open-source tools, most notably the Aequitas bias auditing toolkit.
A Priberam integra a Comunidade IST Spin-Off®.
Os “Priberam Machine Learning Lunch Seminars” são de entrada livre, mediante inscrição prévia.
