
Ang Li is the CEO and cofounder of Simular, the autonomous computer company. Simular builds the full-stack infrastructure for AI agents that use computers like humans do. It was the first to surpass human-level performance on the OSWorld computer-use benchmark with its open-source Agent S framework, which won Best Paper at the ICLR 2025 Agentic AI Workshop. Simular's flagship product, Sai, is a general-purpose autonomous computer in the cloud that operates any software the way a person does.
Ang was formerly a Research Scientist at Google DeepMind, working at the frontier of continual learning, large-scale deep learning, and ML infrastructure. He has published over 50 papers in top AI venues and his work and company have been featured in Forbes, TechCrunch, and VentureBeat. Simular is backed by Felicis and NVentures. His mission: to solve autonomy and enhance lives.
Using publicly available information we constructed an analysis to help you get a feel for this speaker before deciding to attend their session.
Ang Li (ex-Google DeepMind, ex-Baidu Apollo) is CEO and co-founder of Simular, where he builds computer-use agents that operate a real desktop like a human. His team's open-source Agent-S framework reached a state-of-the-art 72.6% on OSWorld with Agent S3 — surpassing reported human-level performance (~72.36%). Expect a research-grounded view, rooted in 'think fast and slow' (System 1 / System 2) agent design, on where autonomous computer-use agents are headed.