Zhanpeng He
Computer Scientist at Columbia University, rlworkgroup
New York City Metropolitan Area United States
Summary
🤩
Rockstar🎓
Top SchoolZhanpeng He is a postdoctoral scholar and robot learning researcher with eight years of experience spanning academia and industry, currently based in Palo Alto and affiliated with Stanford. He holds a PhD from Columbia University in robot learning and computer vision and has held research roles at Amazon and Samsung Research America, bridging cutting-edge research with applied systems. His contributions to open-source robotics toolkits—such as adding Sawyer manipulation environments and task-space control improvements to the popular garage RL framework and developing core World APIs for Metaworld—demonstrate a focus on reproducible multi-task and meta- reinforcement learning benchmarks. Zhanpeng combines strong software engineering from early industry roles with deep RL and perception expertise, enabling robot manipulation research to move toward real-world deployment. Colleagues know him for thoughtful test-driven integrations and careful collision-detection work that noticeably expanded toolkit capabilities beyond standard benchmarks.
8 years of coding experience
Master of Science - MS, Artificial Intelligence, Master of Science - MS, Artificial Intelligence at University of Southern CaliforniaDoctor of Philosophy - PhD, Robot learning and Computer Vision, Doctor of Philosophy - PhD, Robot learning and Computer Vision at Columbia University in the City of New York
English, Chinese
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (10)
algorithm10
openai-gym10
gymnasium10
algorithms10
pytorch10
mujoco10
python10
reinforcement-learning10
api-design9
object-oriented-programming9
Programming languages (4)
Github contributions (5)
Farama-Foundation/Metaworld
Jun 2019 - Dec 2019
Collections of robotics environments geared towards benchmarking multi-task and meta reinforcement learning
Role in this project:
Back-end Developer
Back-end DeveloperContributions:77 commits, 2 PRs, 6 pushes in 6 months
Contributions summary:Zhanpeng primarily contributed to the development of a World API within the repository, suggesting a focus on the core environment logic. Their contributions included adding base classes and APIs for the `multiworld` package, which involved the creation of files for these classes. They also fixed indentation issues, ensuring the code adhered to project standards.
manipulationmulti-taskroboticsbenchmarkingmultiagent-reinforcement-learning
rlworkgroup/garage
Jun 2018 - Jul 2019
A toolkit for reproducible reinforcement learning research.
Role in this project:ML Engineer
ML EngineerContributions:27 commits, 39 PRs, 144 pushes in 1 year
Contributions summary:Zhanpeng significantly contributed to the integration of MuJoCo environments, specifically for the Sawyer robot, within the garage reinforcement learning toolkit. Their work included implementing pick-and-place, reacher, and block stacking environments, along with associated tests. They also focused on task space control and collision detection improvements. These contributions expanded the toolkit's capabilities to encompass robot manipulation tasks.
pytorchreinforcement-learningreproduciblereproducibilityrl-algorithms
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.