Robotics Research Intern, Manipulation

The Team/Project

TRI's Robotics Division develops technologies that enable robots to operate effectively in complex, real-world environments. The project explores how foundation-model agents can understand manipulation tasks, reason about physical environments, and orchestrate reusable robotic capabilities to accomplish diverse objectives. Our research combines vision-language models, robotic perception, motion planning, manipulation, and robot learning. We investigate these questions through hands-on experimentation on physical robotic platforms.

The Challenge

Recent advances in foundation models have significantly improved AI systems' ability to understand visual scenes, reason about tasks, and plan sequences of actions. However, translating these capabilities into reliable physical robot behavior remains challenging. Robots must reason about three-dimensional geometry, physical constraints, contact interactions, and failures during execution.

An important research question is how far foundation-model agents can go by leveraging existing robotic capabilities, and when specialized learned policies become necessary.

The Opportunity

We are seeking a Robotics Research Intern to investigate agentic approaches to robotic manipulation. You will work closely with researchers and engineers to develop systems that connect foundation-model reasoning with physical robot capabilities.

Depending on your background and research interests, your project may focus on spatial reasoning, robotic skill orchestration, execution monitoring, failure recovery, or systematic evaluation of agentic manipulation systems.

This internship offers an opportunity to tackle open research questions while gaining practical experience with cutting-edge AI models and real robotic systems. Your work will contribute to understanding how foundation-model agents can enable more generalizable and reliable robot autonomy.