I am a visiting scientist at the Paul G. Allen School of Computer Science & Engineering, University of Washington. I work at the Personal Robotics Lab under Siddhartha S. Srinivasa. Before this, I received my B.Tech. at IIT Kharagpur, where I was advised by Pabitra Mitra for my thesis project on leveraging prior knowledge for efficient exploration in RL. For the past three years, I have been a part of the Autonomous Ground Vehicle (AGV) Research Group at IIT Kharagpur, where I majorly contributed to the planning and perception modules. This mostly involved working on graph search algorithms, object detection and segmentation problems for operation of autonomous vehicles in unstructured setting. I was also a member of Kharagpur Open Source Society that aims to promote open source through campus wide activities.
My primary research interest is to build safe and sample efficient algorithms to enable robots to learn complex sensorimotor skills in real-world environments. This interest stemmed from my internship at the Computer Vision Center (CVC), Universitat Autònoma de Barcelona, where I worked on unsupervised domain adaptation of driving policies for transfer from simulation to real (sim2real).
Transferring an agent trained in TORCS simulator to CARLA
Adapting to unseen weather conditions for imitation learning agents
Using NDT-Mapping for generating 3D-pointcloud of the IIT-KGP main building
Using rainbow of recent deep learning based techniques for lane segmentation