I am a Research Scientist at Facebook AI Research (FAIR). The goal of my research is to enable robots with the ability to intelligently move (navigate) and do useful things (manipulate) in everyday environments (real world constraints).

These research goals manifest as challenges across the robotics stack ranging from perception to planning. To tackle them I often employ tools such as optimization and deep learning, and sensor modalities like vision and tactile. I also focus on developing 'structured' techniques that can leverage modern machine learning in a way that adds value over traditional robotics methods towards practical robot learning (data efficient, generalizable, interpretable).

Prior to joining FAIR, I received a Ph.D. in Robotics from Georgia Tech where I was advised by Byron Boots and was part of the Institute for Robotics and Intelligent Machines. During my Ph.D. I interned at NVIDIA Research, Amazon Robotics, and Honda Research Institute.

Email: mukadam DOT mh AT gmail DOT com

News:

[2023/05] Tactile Diffusion wins Best Paper Award at ICRA 2023 Workshop on Effective Representations, Abstractions, and Priors for Robot Learning (RAP4Robots)

[2023/05] Two invited talks at ICRA 2023 workshops: Distributed Graph Algorithms for Robotics and 2nd Workshop on Compliant Robot Manipulation

[2023/04] Decentralized Large-Scale Bundle Adjustment accepted to RSS 2023.

[2023/04] Released learning to read braille from touch sensing by closing sim2real gap with tactile diffusion.

[2023/03] Invited talk in the robotics seminar at University of Toronto.

[2023/01] Two neural fields papers accepted to ICRA 2023 on grasping (NGDF) and tactile perception (NCF).

[2022/09] Theseus accepted to NeurIPS 2022.

[2022/09] Two papers on tactile sensing, MidasTouch (oral) and Gravitational pivoting (poster), accepted to CoRL 2022.

[2022/07] Theseus: A library for differentiable nonlinear optimization released, see Meta AI Blog and Twitter thread.

[2022/04] iSDF for real-time neural signed distance fields released. Update: accepted to RSS 2022.

[2022/03] Invited talk in the Robotics Seminar at Cornell.

[2022/01] PatchGraph paper accepted in ICRA 2022.

[2021/12] We open sourced Theseus: A library for differentiable nonlinear optimization (beta release).

[2021/12] Our work on learning energy-based models in optimization (LEO) featured in the CMU ML blog.

[2021/09] Two papers accepted in NeurIPS 2021.

[2021/09] Three papers accepted in CoRL 2021.

[2021/07] Two papers accepted in ICCV 2021.

[2021/06] Paper accepted in IROS 2021.

[2021/06] Habitat 2.0 released and press coverage by TechCrunch, CNET, 20minutes, VentureBeat, FBAI blog.

[2021/06] Our learning robot cinematography work presented at ICRA 2021 and press coverage by VentureBeat, CMU, FBAI twitter.

[2021/03] Two papers accepted in ICRA 2021.

[2020/11] Invited talk in Robot Learning Seminar Series at Mila.

[2020/10] Invited talk in IROS 2020 workshop on Geometric Methods for Robot Learning.

[2020/09] Paper accepted in NeurIPS 2020.

[2020/05] Invited paper on RMPflow accepted in T-ASE.

[2020/03] Organizing RSS 2020 workshop on Good Citizens of Robotics Research.

[2020/03] Organizing RSS 2020 workshop on Structured Approaches to Robot Learning.

[2020/03] Paper accepted in L4DC 2020.

[2020/01] Paper accepted in ICRA 2020.

[2019/09] Two papers accepted in CoRL 2019.

[2019/07] Paper accepted in ISRR 2019.

[2019/07] Successfully defended my Ph.D. thesis.

[2019/06] Paper accepted in IROS 2019.

[2019/06] Our paper on GP motion planning won the IJRR Paper of the Year award! Georgia Tech article here.

[2019/05] I will be joining Facebook AI Research (FAIR) as a Research Scientist this fall.

[2019/05] Patent on Autonomous Vehicle Policy Generation published.

[2019/01] Our work on RMPflow was featured on NVIDIA's Blog and GeekWire.

[2019/01] Paper accepted in ICRA 2019.