Jakub Pachocki

OpenAI (United States)

Papers

1

Total Citations

1,588

H-Index

1

About

Jakub Pachocki is a leading researcher in artificial intelligence and robotics, best known for his pioneering work in reinforcement learning and dexterous manipulation. His most influential contribution, the 2019 paper "Learning dexterous in-hand manipulation" (1,588 citations), revolutionized robotic control by demonstrating that complex, vision-based object reorientation could be learned entirely through simulation. By randomizing physical properties like friction and mass in training, Pachocki’s approach enabled a physical Shadow Dexterous Hand to perform tasks previously thought impossible for machines—such as rotating a block between its fingers—without explicit programming. This work bridged the gap between simulated training and real-world dexterity, setting a new standard for transfer learning in robotics. Beyond this landmark study, Pachocki has advanced large-scale AI systems, contributing to foundational models that push the boundaries of autonomous decision-making. His research has not only garnered thousands of citations but also inspired a generation of roboticists to embrace simulation-based learning. For students and researchers, Pachocki exemplifies how creative use of RL can unlock unprecedented capabilities in embodied AI, making him a pivotal figure in the quest for truly agile, intelligent machines.

Research Focus

Key Achievements

1
H-Index
1
Papers
1,588
Total Citations
1,588
Avg Citations/Paper
🏆 Most Cited Paper
Learning dexterous in-hand manipulation
1,588 citations · 2019
📈 Most Prolific Year: 2019 (1 Papers)
🤝 Key Collaborators: 15
🏛 Institutions: OpenAI (United States)

Top Papers

  1. 1

Key Collaborators

Contact & Links

Available for collaboration
Content generated · 12 days ago