Viraj Joshi
viraj_joshi@utexas.edu

hi! I am a student researcher in Professor Amy Zhang's lab. I completed my undergrad and masters at UT Austin, majoring in computer science and mathematics. I spent my summers as a software engineer at Capital One and Amazon.

I am interested in the techniques to enable generalist agents - whether that be through learning from sim (RL), demonstrations (IL), or a combination of both (offline to online). As of now, I am focused on scaling online, multi-task reinforcement learning for robotics tasks as well as the questions surrounding how to improve the reasoning capabilities of foundation models (RLVR, self-improvement).

email   |   resume   |   github  |  

Unsure how to pronounce my first name? Viraj sounds like V-raj.

A benchmark for massively parallelized multi-task reinforcement learning in manipulation and locomotion.

website adapted from here