# Ang Li

> Ang Li is the co-founder and CEO of Simular, where he is building computer-use agents that can navigate software, complete workflows, and learn from experience. A former Google DeepMind researcher and Baidu Apollo principal scientist with a Ph.D. from the University of Maryland, Li has spent his career on the problem beneath the current agent boom: making machine learning work reliably in environments that keep changing.

- **Role:** CEO and Co-Founder at Simular
- **Organizations:** Simular, Google DeepMind, Baidu Apollo, Facebook AI Research, Carnegie Mellon Robotics Institute, Apple, Google Street View, Comcast Labs DC, University of Maryland, College Park
- **Education:** Ph.D. in Computer Science; research focused on representation learning for visual and linguistic information, University of Maryland, College Park
- **Known for:** Co-founded Simular and raised a $21.5 million Series A in December 2025, following a $5 million seed round., Co-authored Agent S, an open framework for graphical computer use accepted at ICLR 2025., Co-authored Agent S2, a generalist-specialist computer-use framework published at COLM 2025.

## Career timeline

- **2013** — Won first place in the Mid-Atlantic ACM/ICPC regional competition; later competed as an ACM/ICPC World Finalist, placing 39th.
- **2017** — Completed a Ph.D. in computer science at the University of Maryland, College Park.
- **2017-2021** — Led applied research programs at Google DeepMind involving evolutionary neural networks, lifelong learning and Google production systems.
- **2021-2023** — Served as Principal Scientist at Baidu Apollo and led development of continual-learning platforms for autonomous driving.
- **2023** — Co-founded Simular with Jiachen Yang to build computer-use agents and autonomous-computer infrastructure.
- **2024** — Co-authored the Agent S paper and released its open agentic framework for graphical computer use.
- **2025** — Agent S was published at ICLR; Agent S2 followed at COLM, and Simular announced a $21.5 million Series A led by Felicis.
- **2026** — Continued publishing on scaling and reliability for computer-use agents and represented Simular at research and industry events in New York and Seoul.

## Achievements

- Co-founded Simular and raised a $21.5 million Series A in December 2025, following a $5 million seed round.
- Co-authored Agent S, an open framework for graphical computer use accepted at ICLR 2025.
- Co-authored Agent S2, a generalist-specialist computer-use framework published at COLM 2025.
- Published more than 39 AI research papers, according to Simular's official biography.
- Served on senior program committees for AAAI and IJCAI and as an area chair for NeurIPS meta-learning workshops.
- Won the 2013 Mid-Atlantic ACM/ICPC regional and placed 39th as an ACM/ICPC World Finalist.
- Led applied research programs at Google DeepMind from 2017 to 2021 and continual-learning platform work at Baidu Apollo from 2021 to 2023.

## Latest updates

- **2026-07** — Visited Seoul during ICML week, hosted a computer-use-agent researcher roundtable and discussed Korea's potential role in supplying hardware for autonomous agents.
- **2026-06** — Appeared in a Simular-posted interview from the NYSE studio discussing Sai, reliability and small-business use cases.
- **2026-04** — Co-authored 'On the Reliability of Computer Use Agents,' extending his research focus on dependable agent execution.
- **2026-01** — Appeared on Fund/Build/Scale to discuss founder decision-making, small teams and the instability of defensibility in AI.
- **2025-12** — Simular announced a $21.5 million Series A led by Felicis and debuted Simular 1.0.

## Links

- Website: https://angli.ai/
- LinkedIn: https://www.linkedin.com/in/angli-ai
- Twitter/X: https://x.com/angli_ai
- GitHub: https://github.com/simular-ai/Agent-S

---

Profile page: https://yespress.io/ang-li
Published by YesPress — https://yespress.io
Last updated: 2026-08-17