Human Interactive Driving Intern – World Models

Job description

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 advancing the state of the art in AI, robotics, driving, and material sciences. This is a paid 12-week internship opportunity and is a hybrid, in-office role. [Here’s a glimpse into the Internship experience from some of our TRI interns!](https://www.tri.global/tri-internship-program) **The Team** You’ll be joining a multidisciplinary research team focused on developing a world foundation model for driving—a unified, transferable representation of driving knowledge built from large-scale real-world and simulated data. Together, we are tackling the complex challenges of multi-agent interaction, causal reasoning, and embodied intelligence in highly dynamic, real-world driving environments. Our approach is inspired by cutting-edge generative modeling techniques and powered by scalable infrastructure and open-source data, and we work closely with engineers to deploy our models in simulation and on hardware. **The Internship** As a Ph.D. Research Intern, you will conduct original research at the intersection of machine learning and autonomous driving. Your work may focus on developing novel components of world models, improving decision-making through model-based RL, advancing 3D perception for dynamic scenes, or enhancing simulation-to-reality transfer. You’ll have the opportunity to prototype and evaluate your ideas in closed-loop simulations and contribute to research publications alongside TRI scientists. This is an opportunity to test and refine your research in a collaborative, high-impact environment, while working with real-world data and contributing to the future of autonomous systems. The internship will be in our headquarters and include competitive compensation, befitting the challenging but fun nature of the research work at TRI. Applicants with relevant publications in the fields above and good collaboration skills are highly encouraged to apply. ### Responsibilities - Conduct original research in one or more areas: world modeling, multi-agent interaction, reinforcement learning, perception, or simulation-to-reality transfer. - Collaborate closely with full-time researchers on the design, training, and evaluation of learning-based driving systems. - Contribute to building and experimenting with task-aware, multi-modal, and uncertainty-aware models. - Develop and evaluate prototypes in closed-loop simulation environments and, time permitting, on high-performance autonomous driving hardware. - Present research findings through internal talks and work towards a top-tier academic publication. - Integrate and work with large-scale datasets (open-source and internal). ### Qualifications - Currently enrolled in a Ph.D. program in Computer Science, Robotics, Machine Learning, or a related field. - Strong background in machine learning, particularly in areas such as deep learning, generative models, reinforcement learning, or probabilistic modeling. - **Demonstrated experience with one or more of the following:** World models (e.g., latent dynamics, diffusion-based models), Model-based RL or decision-making, 3D perception or sensor fusion, and Large-scale simulation for robotics or autonomous systems. - Prior publication(s) in top-tier conferences (NeurIPS, ICLR, ICML, CVPR, ICRA, CoRL, etc.) - Proficiency with Python and PyTorch. - Familiarity with AWS services (S3, EC2, and SageMaker) and open-source driving datasets (nuScenes, Waymo, Argoverse, etc.) is a plus.