School of Interactive Computing
Georgia Institute of Technology
Email varunagrawal @ gatech.edu
I am a recent PhD graduate from the School of Interactive Computing at Georgia Tech, where I was advised by Prof. Frank Dellaert.
My research goal is to build robotic systems that reason probabilistically under uncertainty over geometry to enable efficient and reliable autonomy in the real world.
Currently, I am focused on proprioceptive state estimation of legged robots using concepts from machine learning and nonlinear control. Robot dynamics are intuitive constraints since they obey nature's laws, thus being able to model the observable and the latent processes underlying legged locomotion, we can extract useful information about the robot's state without the need for exterioceptive sensors such as cameras.
I am also a maintainer of GTSAM, an industrial-strength factor graph library for robotics, as well as its sister projects (e.g. GTDynamics).
Latest Updates
- Defended my PhD dissertation on 17th September 2025.
- Presented a poster on RepoAuditor at US-RSE 2025, my summer project with the Georgia Tech Center for Scientific Software Engineering.
- Our paper on a new state estimation metric has been accepted to ICRA 2024.
- 2 papers accepted to ICRA 2023!
- Paper on Legged Robot State Estimation accepted to Humanoids 2022.
Teaching
- OMSCS 6475 Computational Photography, TA with Prof. Irfan Essa
- CS 4476/6476 Fall 2017 Computer Vision, TA with Prof. James Hays
- CS 4476/6476 Fall 2016 Computer Vision, TA with Prof. James Hays