Pieter Abbeel
Pieter Abbeel studies how machines learn to act through demonstrations, feedback, and experience across tasks. He completed his Stanford doctorate with Andrew Ng in 2008 and became a Berkeley professor. His research includes helicopter control and robotic laundry folding, with collaborators and students. He coauthored Trust Region Policy Optimization and, with Chelsea Finn and Sergey Levine, Model-Agnostic Meta-Learning. He also co-founded Covariant with Peter Chen, Rocky Duan, and Tianhao Zhang to develop robot-learning systems for warehouse tasks. In 2024, Amazon licensed Covariant models non-exclusively and hired part of its team. Abbeel is an Amazon Scholar and Frontier AI and Robotics co-leader, alongside his Berkeley role. His industrial work now connects frontier models, agents, and robotics.
- Now at
- University of California, Berkeley and Amazon
- Works
- 621
Publication numbers from OpenAlex, as of 14 October 2026. Citations and h-index are hidden because this OpenAlex record may be incomplete or mix in other authors.
Timeline
-
2008
Stanford doctorate
Abbeel completes his computer science doctorate with Andrew Ng as adviser.
-
2010
Learning physical tasks
His work with colleagues includes helicopter flight and robotic laundry folding.
-
2015–2017
Learning methods
He coauthors Trust Region Policy Optimization and Model-Agnostic Meta-Learning with their teams.
-
Before August 2024
Covariant
Abbeel, Peter Chen, Rocky Duan, and Tianhao Zhang co-found the robot-learning company.
-
2024
Amazon research
A model license and staff move precede a new San Francisco AI lab announced with David Luan.
-
October 2026
Models and robotics
His Amazon role connects him with frontier model research and continuing robotics work.
Sources and credits
Photo: Pieterabbeel, via Wikimedia Commons. License: CC BY-SA 4.0. Source: commons.wikimedia.org. Cropped to a square and resized. Ask us to remove or replace this photo.
