Robotics Intern, Dexterous Manipulation – Continual Robot Learning (CRL)

At Toyota Research Institute (TRI), we’re on a mission to improve the quality of human life. We’re developing new tools and capabilities to amplify the human experience. To lead this transformative shift in mobility, we’ve built a world-class team in Human-Centered AI, Human Interactive Driving, Energy and Materials, Machine Learning, and Robotics.

This is a Summer 2023 paid 12-week internship opportunity. Please note that this internship will be a hybrid in-office role.

The Mission

We conduct research aimed at creating general-purpose robots that can scale to complex tasks in diverse environments. Our belief is that these robots must increase in capability and behavioral complexity over time without requiring impractical amounts of human expertise and engineering. To accomplish this, we’re investigating how learning can be used to automatically combine existing behavioral primitives (skills) in new ways to recover from unexpected failures and solve novel tasks.

The Team

The Dexterous Manipulation Team’s charter is to push the frontiers of research in robotics and machine learning to develop the future capabilities required for general-purpose robots able to operate in unstructured environments such as homes. We currently focus more on interaction and manipulation than we do on locomotion and much of our work currently takes place on bimanual robot manipulation stations. A long-running goal of our group is to lay the research groundwork for general-purpose in-home robots that can assist humans with tasks such as preparing food, keeping their house clean, or doing laundry.

The Internship

We’re looking for an intern to help us bridge the gap in robotics between discrete symbolic planning and continuous low-level skills! While (policy) learning has increasingly played a key role in short-horizon behaviors such as object grasping, complex long-horizon tasks (such as a robot cooking a hamburger) still rely on manual behavior sequencing or explicitly-designed symbolic planning. We want to bridge this gap and allow robots to automatically sequence and compose known short-horizon behaviors to accomplish complicated tasks without requiring significant human engineering; the goal is to create a robot capable of reusing existing behavior in novel ways when encountered with new situations (such as a novel task or an unexpected edge case).

The intern who joins our team will be expected to create working code prototypes, interact frequently with team members, run experiments with both simulated and real (physical) robots, and participate in publishing the work to peer-reviewed venues.

Qualifications (While not strictly mandatory, a subset of the following is preferred)

• Experience with task and motion planning (TAMP).
• Experience with symbolic reasoning, task planning, and/or search.
• Experience with hierarchical RL and/or offline RL.
• Experience with machine learning and familiarity with a major deep-learning framework such as PyTorch.
• Familiarity with computer vision, particularly visual scene/object embeddings.
• Familiarity with NLP and visual/language embeddings.
• Strong software development skills in Python (C++ is also useful).
• A passion for robotics and doing research grounded in important fundamental problems.

Please reference this Candidate Privacy Notice to inform you of the categories of personal information that we collect from individuals who inquire about and/or apply to work for Toyota Research Institute, Inc. or its subsidiaries, including Toyota A.I. Ventures GP, L.P., and the purposes for which we use such personal information.

TRI provides Equal Employment Opportunity without regard to the applicant’s race, color, creed, gender, gender identity or expression, sexual orientation, national origin, age, physical or mental disability, medical condition, religion, marital status, genetic information, veteran status, or any other status protected under federal, state or local laws.