Pulkit Agrawal

Pulkit Agrawal is an Associate Professor in the Department of Electrical Engineering and Computer Science (EECS) at the Massachusetts Institute of Technology (MIT). His research interests include Artificial Intelligence + Machine Learning, Graphics and Vision, and Robotics. Agrawal's work has been recognized with awards such as the IROS Toshio Fukuda Young Professional Award for contributions to robot learning and sim-to-real policy learning. He is affiliated with the MIT EECS Communication Lab and maintains a profile page at https://www.eecs.mit.edu/people/pulkit-agrawal/.

Pulkit Agrawal is a researcher at the Massachusetts Institute of Technology, focusing on Robotics, Computer Vision, Artificial Intelligence, and Reinforcement Learning. His work explores curiosity-driven exploration through self-supervised prediction, automated echocardiogram interpretation, and human pose estimation with iterative error feedback. Agrawal investigates transfer learning on ImageNet, conditional generative modeling for decision-making, and visual predictive models for physics-based tasks like billiards. His research also includes embodied question answering, in-hand object re-orientation, and offline reinforcement learning. Agrawal emphasizes intuitive physics learning and visual dexterity for complex object manipulation, aiming to enhance robotic autonomy and generalization.