Robotics Algorithms · Internship

Robot Manipulation Algorithm Intern

Responsibilities

Develop stable, natural mappings from human hand motion capture, keypoints, and pose estimates to dexterous robot hands.
Implement, train, and iterate imitation-learning methods such as ACT, Diffusion Policy, and vision-language-action models.
Combine modern models with high-dimensional tactile information; design experiments and improve real-robot success and generalization.

Currently pursuing a master’s or PhD in computer science, robotics, control, or a related field.
Proficient in Python and PyTorch, with good coding standards and reproducible experiment practices.
Familiar with ROS or ROS 2 and practical robot deployment; dexterous-hand experiments are preferred.
Understand major imitation-learning methods and independently train, debug, and analyze models.
Enjoy hands-on robotics, engineering problem-solving, and running algorithms on real hardware.

Experience with dexterous hands, robot arms, or human-hand teleoperation.
Robotics/AI competition achievements or manipulation publications.
Experience setting up Isaac Gym, Isaac Sim/Lab, or MuJoCo environments.
Reinforcement-learning projects or tactile, visual, and action multimodal fusion.

Dexterous hands, human-hand retargeting, and tactile-enhanced imitation learning.
Real arms, hands, vision/tactile sensors, and Ego acquisition platforms.
The full R&D cycle from algorithms and data to training and real-robot deployment.
Collaboration across tactile sensing, dexterous manipulation, and embodied AI teams.