Sriram Yenamandra

I am a first year Computer Science PhD student at Stanford University. I completed my Master's degree in Computer Science (Machine Learning specialization) from Georgia Tech. During my Master's degree, I worked in Prof. Dhruv Batra's lab on solving embodied mobile manipulation tasks. At Georgia Tech, I also had the privilege of being advised by Prof. Judy Hoffman on the problems of visual domain adaptation and bias identification.

Before coming to Georgia Tech, I earned my bachelor's degree in Computer Science and Engineering with a minor in Applied Statistics from IIT Bombay, where I worked on the problem of image inpainting under the supervision of Prof. Suyash Awate.

Email  /  CV  /  Google Scholar  /  Twitter  /  Github

My research interests broadly lie in computer vision, machine learning and robotics. My goal is to develop general-purpose robotic assistants that can quickly learn to perform new tasks in novel environments. The development of such generalist agents remains challenging, particularly due to scarcity in real robot data and the lack of realism in simulated environments. I am interested in utilizing multiple synthetic data sources (eg. simulations, diffusion models) and uncurated video datasets to solve real-world (video, 3D or embodied) tasks. My previous work has contributed to this objective by developing a framework for using simulations to build real-world mobile manipulators, tackling visual domain shifts (such as sim2real), and employing generative models to pinpoint data slices that may need additional collection.

(Design and CSS courtesy: Jon Barron)