Research Intern: Intelligent Multimodal Representation for Dexterous Manipulation

Honda Research Institute USA; San Jose (CA); On-site; USD 34,000 - 48,000; Full time

Job summary

Honda Research Institute USA in San Jose, CA invites a Research Intern to advance multimodal perception and learning for dexterous manipulation using multi-fingered robot hands. The role emphasizes contact-rich manipulation under partial observability, with both simulation and real hardware for evaluating learning algorithms.

The intern will develop and test algorithms, process multimodal data, and contribute to publications and patents where applicable.

Qualifications

• PhD or MS candidate in robotics, CS, or related field.
• Experience processing and learning from multimodal sensory data (tactile, force, RGB-D).
• Experience implementing and evaluating learned policies on real robots.
• Experience with robotic simulators (e.g., Isaac Gym or MuJoCo).
• Proficient with Python, C++, and PyTorch.
• Experience with ROS 2.

Responsibilities

• Develop and evaluate learning algorithms for multimodal perception and dexterous manipulation.
• Collect and process multimodal sensory data and conduct representation and policy training.
• Integrate and validate learned models on robotic manipulation platforms.
• Implement baselines and perform benchmark evaluations on hardware and in simulation.
• Document findings and contribute to internal research reports.
• Contribute to patents and publish results at top-tier conferences and journals.

Skills

Multimodal perception

Dexterous manipulation

Python

C++

PyTorch

ROS 2

Data processing

Education

PhD candidate in robotics

MS candidate in robotics

Tools

Isaac Gym

MuJoCo

ROS 2

Job description

Research Intern: Intelligent Multimodal Representation for Dexterous Manipulation - Honda Research Institute USA

Job Number: P26INT-24

Honda Research Institute USA (HRI-US) is seeking a self-motivated research intern to advance multimodal perception and learning for dexterous manipulation using multi-fingered robot hands. The internship project focuses on contact-rich manipulation under partial observability, exploring how multimodal sensing can inform both manipulation actions and active exploration to acquire relevant information. The successful candidate will utilize both simulation and real hardware to implement, evaluate, and iteratively improve learning algorithms.

San Jose, CA

Key Responsibilities

• Develop and evaluate learning algorithms for multimodal perception and dexterous manipulation.
• Collect and process multimodal sensory data and conduct representation and policy training.
• Integrate and validate learned models on robotic manipulation platforms.
• Implement baselines and perform benchmark evaluations on hardware and in simulation.
• Document findings and contribute to internal research reports.
• Contribute to the portfolio of patents and publish research results when applicable at top-tier conferences and journals in robotics and machine learning.

Minimum Qualifications

• Ph.D. or highly qualified M.S. candidate in robotics, computer science, or a related field.
• Experience processing and learning from multimodal sensory data (tactile, force, RGB-D, etc.).
• Experience implementing and evaluating learned policies on real-world robots.
• Experience with robotic simulators (e.g., Isaac Gym or MuJoCo).
• Proficient with Python, C++, and PyTorch.
• Experience with the Robot Operating System (ROS 2).

Bonus Qualifications

• Experience with multimodal representation learning and sensor fusion.
• Experience with learned world models and predictive modeling.
• Experience with vision-language-action or robot foundation models.
• Experience with dexterous manipulation and contact dynamics.
• Experience with sim-to-real or reinforcement learning on hardware.

Years of Work Experience Required 0

Desired Start Date May 2027

Internship Duration 3 Months

Position Keywords Dexterous Manipulation, Multimodal Representation, Active Exploration