Deep Learning for Dexterous Manipulation - Honda Research Institute USA

Deep Learning for Dexterous Manipulation

Your application is being processed

Deep Learning for Dexterous Manipulation

Job Number: P24INT-06
​The main objective of this research is to advance the manipulation intelligence of multi-fingered robot hands. We explore a hierarchical approach that combines large and small models to achieve human-like in-hand dexterous manipulation, addressing challenges such as long task horizons and partial observability. The ideal candidate will utilize both simulation and real hardware to implement, evaluate, and iteratively improve the learned policy.
San Jose, CA

 

Key Responsibilities

 
  • Develop learning methods for dexterous in-hand object manipulation using multi-fingered robot hands.
  • Data collection and policy training.
  • Implement baselines and perform benchmark evaluation for the proposed policy in simulation and on hardware.
  • Contribute to the creation and evaluation of various related technologies.
  • Contribute to academic publications for top-tier conferences and journals in robotics and machine learning.

 

Minimum Qualifications

 
  • Ph.D. or highly qualified M.S. candidate in robotics, computer science, mechanical engineering, or a related field.
  • Proficient with robot learning for manipulation (behavior cloning and reinforcement learning).
  • Experience with auditory or visual time-sequence modeling.
  • Experience with policy deployment on real-world robots.
  • Experience with robotic simulators (e.g., Isaac Gym or MuJoCo).
  • Proficient with Python and C++.
  • Proficient with PyTorch.
  • Experience with Robot Operating System (ROS).

 

Bonus Qualifications

  • Experience with large vision-language-action models or vision-language models.
  • Experience with in-hand dexterous manipulation on multi-fingered robot hands.
  • Experience with representation learning using RGB, RGB-D, and tactile data as inputs.
  • Knowledge in contact dynamics and contact mode switching.
  • Experience in online reinforcement learning with hardware robots.
  • Experience with Sim2Real approaches. 

 

Years of Work Experience Required  0
Desired Start Date 5/19/2025
Internship Duration 3 Months
Position Keywords In-hand Object Manipulation, Time-sequence Deep Learning, Robotics

 

  

Alternate Way to Apply

Send an e-mail to careers@honda-ri.com with the following:
- Subject line including the job number(s) you are applying for 
- Recent CV 
- A cover letter highlighting relevant background (Optional)

Please, do not contact our office to inquiry about your application status.