Covariant
OpenAI
KU Leuven
Stanford University
Sergey LevineAssociate Professor · UC BerkeleyResearcher↗
Researcher · Academic · Founder
Pieter Abbeel
AmazonDistinguished Scientist, VP, Scholar
UC BerkeleyProfessor
Born 1977
Abbeel helped robots learn from demonstrations and trial and error, including handling objects and learning new skills from few examples.[2] He co-founded Covariant, bringing robot learning into warehouses, and Gradescope, which helps teachers grade student work. His research also helped develop MAML, a method for learning new tasks quickly.[1][3]
Career & education
Current rolesAmazon2024-presentDistinguished Scientist, VP, ScholarUC Berkeley2008-presentProfessor
Amazon2024-presentUC Berkeley2008-presentEducation
Work and education history is primarily focused on AI-relevant roles and may not be comprehensive.
Citation Trend
Total citations to scholarly work, including research papers, technical reports, books and theses, as recorded by Google Scholar. Each point is a dated snapshot.
citations298,103As of Nov 1, '26
citations+9,119Since last snapshot, Oct 1, '26
Selected contributions
Denoising Diffusion Probabilistic Models
With Jonathan Ho and Ajay JainGenerated images by learning to turn random noise into a picture, step by step.
Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks
With Chelsea Finn and Sergey LevineMAML trains models to adapt to new tasks with only a few examples.
References
- UC Berkeley — Pieter Abbeel, robot learning and named contributionsvcresearch.berkeley.edu
Abbeel helped robots learn from demonstrations and trial and error, including handling objects and learning new skills from few examples. He co-founded Covariant, bringing robot learning into warehouses, and Gradescope, which helps teachers grade student work. His research also helped develop MAML, a method for learning new tasks quickly.
- UC Berkeley — learning from demonstrations and trial and errorvcresearch.berkeley.edu
Abbeel helped robots learn from demonstrations and trial and error, including handling objects and learning new skills from few examples.
- Model-Agnostic Meta-Learning — original paperarxiv.org
His research also helped develop MAML, a method for learning new tasks quickly.
Based on public sources. If accuracy is important, verify details using the links or references.
For corrections, updates, or requests to remove information, contact us.
Reviewed July 23, 2026 · Checked for accuracy and new information.
