Robotics Machine Learning Research Scientist - Large Behavior Models

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 ground-breaking shift in mobility, we’ve built an extraordinary team in Automated Driving, Energy & Materials, Human-Centered AI, Human-Interactive Driving, and Robotics.

The Mission

We are working to create general-purpose robots capable of accomplishing a wide variety of dexterous tasks. To do this, we're building general-purpose machine learning foundation models for dexterous robot manipulation. These models, which we call Large Behavior Models (LBMs), use generative AI techniques to produce robot action from sensor data and human request. To accomplish this, we are creating a large curriculum of embodied robot demonstration data and combining that data with a rich corpus of internet-scale text, image, and video data. We are also utilizing high-quality simulation to augment real-world robot data with procedurally generated synthetic demonstrations.

The Team

The Robotics Machine Learning 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 or factories.

The Job

We have several research thrusts under our broad mission, and we are looking for a research scientist in any of these areas:

-Data-efficient and general algorithms for learning robust policies using multiple sensing modalities: proprioception, images, 3D representations, force, and dense tactile sensing.

-Scaling learning approaches to large-scale models trained on diverse sources of data, including web-scale text, images, and video.

-Leveraging test time computation for embodied applications.

-Quick and efficient improvement of learned policies.

-Continual Learning and Adaption

-Multi-Modal Reasoning Models.

-Structured hierarchical reasoning using learned models.

-Reinforcement Learning with Language Action Models

-Leveraging history and memory for learning policies for long context tasks.

-Improving robustness and few-shot generalization by using sub-optimal and self-play data.

-Interactive agents that can reduce the embodied and instructional ambiguity and can seek help and clarification.

The researcher who joins will be encouraged to collaborate in our code infrastructure, work together with team members, run experiments with both simulated and real (physical) robots, and participate in publishing work to peer-reviewed venues and open-sourcing code. We’re looking for a research scientist who is comfortable working with both existing large static datasets as well as a growing dynamic corpus of robot data.

Qualifications

• Hands-on experience with using machine learning for learned control, including RL, offline RL or behavior cloning, for manipulation. Or experience with machine learning and familiarity with large multi-modal datasets and models.
• Strong software development skills in Python.
• A “make it happen” demeanour and comfort with fast prototyping.
• A passion for robotics and doing research grounded in important fundamental problems.
• A track record of relevant publications in top international conferences (RSS, NeuRIPS, ICML, ICLR, CoRL, ICRA, IROS, …)

Bonus Qualifications

• Hardware experience.
• A track record of relevant open-source software contributions
• Experience working with large-scale datasets and multi-node training

The pay range for this position at commencement of employment is expected to be between $176,000 and $264,000/year for California-based roles. Base pay offered will depend on multiple individualized factors, including, but not limited to, business or organizational needs, market location, job-related knowledge, skills, and experience. TRI offers a generous benefits package including medical, dental, and vision insurance, 401(k) eligibility, paid time off benefits (including vacation, sick time, and parental leave), and an annual cash bonus structure. Additional details regarding these benefit plans will be provided if an employee receives an offer of employment.

Please reference thisCandidate Privacy Noticeto 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 is fueled by a diverse and inclusive community of people with unique backgrounds, education and life experiences. We are dedicated to fostering an innovative and collaborative environment by living the values that are an essential part of our culture. We believe diversity makes us stronger and are proud to provide Equal Employment Opportunity for all, without regard to an 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.

It is unlawful in Massachusetts to require or administer a lie detector test as a condition of employment or continued employment. An employer who violates this law shall be subject to criminal penalties and civil liability. Pursuant to the San Francisco Fair Chance Ordinance, we will consider qualified applicants with arrest and conviction records for employment.