Hejia Zhang
Staff Research Scientist, AI Robotics at NIO
San Francisco Bay Area United States
Summary
🤩
Rockstar🎓
Top SchoolHejia Zhang is a staff research scientist specializing in dexterous, human-like robot manipulation with nine years of experience bridging academic research and industry productization. Currently leading manipulation research at NIO USA after roles at Skild AI and CMU, she brings deep RL and robotics expertise from a PhD in computer science at USC and a bioengineering BS from Zhejiang University. Her contributions to open-source toolkits like rlworkgroup/garage highlight practical work integrating RL algorithms with ROS, Sawyer robot environments, and safety checks via MoveIt!, enabling reproducible robotics experiments. Hejia’s background spans both bioengineering and computer science, which informs a multidisciplinary approach to contact-rich manipulation and robust control. Colleagues describe her as a hands-on researcher who moves ideas from simulation to hardware while keeping production safety and reproducibility front and center.
9 years of coding experience
6 years of employment as a software developer
Doctor of Philosophy - PhD (M.S. completed 2020) Computer Science, Doctor of Philosophy - PhD (M.S. completed 2020) Computer Science at University of Southern CaliforniaBachelor of Science - BS Bioengineering, Bachelor of Science - BS Bioengineering at Zhejiang University
Github Skills (8)
algorithm10
pytorch10
robotics10
python10
ros10
tensorflow9
openai-gym7
gymnasium7
Programming languages (10)
Github contributions (5)
rlworkgroup/garage
Jun 2018 - Nov 2018
A toolkit for reproducible reinforcement learning research.
Role in this project:
ML Engineer
ML EngineerContributions:74 commits, 19 PRs, 155 pushes in 5 months
Contributions summary:Hejia primarily contributed to the development and integration of reinforcement learning algorithms within the project. Their work involved refactoring existing code related to ROS integration, adding new environments and functionalities for the Sawyer robot, and implementing safety checks using MoveIt!. The user focused on integrating the developed algorithms, utilizing PyTorch and potentially TensorFlow, as suggested by the repository topics. This work is geared toward enabling reproducible RL research on robotics platforms.
pytorchreinforcement-learningreproduciblereproducibilityrl-algorithms
icaros-usc/milp_constrained_gan
Apr 2020 - Oct 2020
Using MILP to encode constraints of generated game levels.
Contributions:71 commits, 1 push in 5 months
levelsmilpencode
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.