Anish Bhattacharya
Anish Bhattacharya
Embedded Files
Vision and learning for sim-to-real robots
Vision and learning for sim-to-real robots
Seeking full-time roles starting Spring/Summer 2026!
Seeking full-time roles starting Spring/Summer 2026!
Some related topics: multimodal perception, policy adaptation from simulation or large data corpora, leveraging uncertainties of large models, and learned navigation policies.
Some related topics: multimodal perception, policy adaptation from simulation or large data corpora, leveraging uncertainties of large models, and learned navigation policies.
I am a PhD student at GRASP with Nikolai Matni and Vijay Kumar, and am  grateful to be supported by the NSF GRFP. I've interned at Microsoft Research, Zoox, and RPG Lab (Davide Scaramuzza).
I am a PhD student at GRASP with Nikolai Matni and Vijay Kumar, and am  grateful to be supported by the NSF GRFP. I've interned at Microsoft Research, Zoox, and RPG Lab (Davide Scaramuzza).
My work focuses on enabling faster vision-based robot control through the following two thrusts:
My work focuses on enabling faster vision-based robot control through the following two thrusts:
Developing models and training frameworks: running small and fast vision transformers onboard a quadrotor for vision-based control, and enabling sim-to-real transfer of a vision-based flight policy with a novel post-training framework.
Neural implicit models for asynchronous perception: using neural-based rendering models to generate novel-view event camera datastreams, and using properties of event cameras to enable events-to-control via reinforcement learning or modeling via Neural ODEs.
anish1 [at] seas.upenn.edu
anish1 [at] seas.upenn.edu
Google Sites
Report abuse