Robotics Intern, World Models, Multimodal Sensing
Toyota Research Institute
Summary
We are working to create general-purpose robots capable of accomplishing a wide variety of dexterous tasks. To do this, our team is building general-purpose machine learning foundation models for dexterous robot manipulation. These models, which we call Large Behavior Models, use generative AI techniques to produce robot actions from sensor data and human requests. 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 using high-quality simulation to augment real-world robot data with procedurally-generated synthetic demonstrations.
Job description
The Mission
We are working to create general-purpose robots capable of accomplishing a wide variety of dexterous tasks. To do this, our team is building general-purpose machine learning foundation models for dexterous robot manipulation. These models, which we call Large Behavior Models, use generative AI techniques to produce robot actions from sensor data and human requests. 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 using high-quality simulation to augment real-world robot data with procedurally-generated synthetic demonstrations.
The Team
The Large Behavior Models team 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!
The Internship
We are looking for a research intern in the area of data-efficient and general algorithms for learning robust policies leveraging multiple sensing modalities, including proprioception, images, force, audio, and dense tactile sensing. The intern who joins our team will be expected to create working code and/or hardware 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. We're looking for an intern who is comfortable working with both existing large static datasets as well as a growing and dynamic corpus of robot data.
Qualifications
- Strong hands-on experience with machine learning and familiarity with large datasets and generative models.
- Research experience in world models, action-conditioned prediction, or learned control, with an interest in contact-rich manipulation.
- Strong software development skills in Python and experience with modern deep learning frameworks such as PyTorch.
- A "make it happen" attitude and comfort with rapid prototyping and testing in simulation and on physical robots.
- A passion for learning how robot actions shape physical interactions, including contact dynamics, deformation, slip, and force transmission.
- A creative approach to world action models that integrate tactile and force sensing with vision and proprioception to predict interaction outcomes and guide robot behavior.
- Experience with tactile sensors, force/torque sensing, contact simulation, or force-controlled manipulation is strongly preferred.
Requirements
- Strong hands-on experience with machine learning and familiarity with large datasets and generative models
- Research experience in world models, action-conditioned prediction, or learned control, with an interest in contact-rich manipulation
- Strong software development skills in Python and experience with modern deep learning frameworks such as PyTorch
- A "make it happen" attitude and comfort with rapid prototyping and testing in simulation and on physical robots
- A passion for learning how robot actions shape physical interactions, including contact dynamics, deformation, slip, and force transmission
- A creative approach to world action models that integrate tactile and force sensing with vision and proprioception to predict interaction outcomes and guide robot behavior
- Experience with tactile sensors, force/torque sensing, contact simulation, or force-controlled manipulation is strongly preferred
Benefits
- Medical, Dental & Vision
- Paid Time Off
- Holiday Pay
- Sick Time
Company overview
Toyota Research Institute (TRI) is a premier research and development subsidiary of Toyota Motor Corporation, established in 2016 with a $1 billion investment over five years. Headquartered in Los Altos, California with additional offices in Cambridge, Massachusetts, TRI focuses on advancing technologies in artificial intelligence, robotics, automated driving, and materials science. Led by CEO Dr. Gill Pratt, a renowned roboticist and former DARPA official, the institute employs 201-500 highly skilled researchers, over half of whom hold PhD degrees. TRI's mission is to amplify human ability and improve quality of life through technological innovation, with specific research areas including human-centered AI, energy and materials science, human interactive driving, and robotics. The institute collaborates with academic institutions and industry partners to develop cutting-edge solutions for mobility challenges, battery technology, and sustainable energy systems. TRI operates as an independent research entity within the Toyota ecosystem, maintaining a startup-like culture while benefiting from Toyota's substantial resources and global automotive expertise.