Internship in Robotics, Dexterous Handling - Extensive Behavioral Models

At the Toyota Research Institute (TRI), we are on a mission to enhance the quality of human life. We are creating new tools and abilities to enhance the human experience. Our world-class team leads this transformative shift in mobility, with expertise in Human-Centered AI, Human Interactive Driving, Energy and Materials, Machine Learning, and Robotics! This is a paid 12-week internship opportunity for summer 2024. Please note that this internship requires an in-office presence. Our Mission We are working to create universal robots capable of performing a broad range of skilled tasks. Our team is developing universal machine learning fundamental models for skilled robot manipulation, known as Large Behavior Models. These models employ generative AI techniques to draw robot action from sensor data and human instruction. We are creating an extensive curriculum of embodied robot demonstration data and merging it with a rich body of internet-scale text, image, and video data. We are also using top-tier simulation to enhance real-world robot data with procedurally generated synthetic demonstrations. The Team The Dexterous Manipulation Team aims to advance research in robotics and machine learning to develop needed future capabilities for universal robots who can function in unstructured environments, such as homes! The Internship Our broad mission includes several research thrusts, and we are seeking a research intern for any of these areas: - Developing data-efficient and universal algorithms for learning robust policies using multiple sensing modalities: proprioception, images, force, and dense tactile sensing. - Scaling learning approaches to large-scale models trained on diverse data sources, including web-scale text, images, and videos. - Quick and efficient enhancement of learned policies. The intern joining our team is expected to create operational code prototypes, engage frequently with team members, run experiments with both simulated and real-life robots, and contribute to the publication of the work in peer-reviewed outlets. We are in search of an intern who can comfortably work with both existing large static datasets and a growing, dynamic corpus of robot data. Qualifications - Either hands-on experience with using machine learning for learned control, including RL, offline RL or behavior cloning, for manipulation. Hardware experience is not required but strongly preferred. -OR- Strong experience with machine learning and familiarity with large datasets and models. - Strong software development skills in Python. - A “make it happen” attitude and comfort with quick prototyping. - A passion for robotics and conducting research grounded in vital fundamental problems. Please refer to this Candidate Privacy Notice to inform you about the categories of personal information 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 driven by a diverse and inclusive community of people with a variety of backgrounds, education and life experiences. We are committed to creating an innovative and collaborative environment by embodying the values that form an essential part of our culture. We believe diversity makes us stronger and are proud to provide Equal Employment Opportunity for all, regardless of 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 protected status under federal, state, or local law.