Sudharshan Suresh

Sudharshan Suresh

••••••• [at] gmail [dot] com

(click to unscramble)

I'm a staff research scientist and technical lead at Boston Dynamics on the Atlas humanoid project. I currently focus on scaling humanoid policies through alternative data sources like egocentric human data.

I earned my Ph.D. in Robotics from CMU, while working as a visiting researcher at FAIR (Meta) on the dexterous manipulation team. My thesis enabled robots to learn from vision and touch, see popular work like NeuralFeels and MidasTouch.

Boston Dynamics Icon

News

2025 ↓

Jun

I did a deep-dive into our whole-body manipulation and vision stack at the RSS Dexterous Manipulation Workshop, link to the invited talk here.

May

I'm a featured interview in our team's video release and a contributing author to the accompanying blogpost

May

Selected to be part of the RSS Pioneers Workshop, one of 33 PhD and early-career industry researchers.

2024 ↓

Nov

NeuralFeels is published in Science Robotics and featured as cover of November issue (spotlight article).

Oct

My work on Atlas was featured in their autonomous demo (IEEE, TechCrunch, Verge).

Work

Boston Dynamics Atlas Video Release

RSS Dexterous Manipulation Workshop Talk - Making Atlas See the World

Atlas Data Visualization

I'm a research scientist and technical lead on the Atlas VLA team. I currently focus on scaling humanoid policies through alternative data sources like egocentric human data (more details soon). My work is often split between ML training, robot deployment, and coordinating data collection efforts. Previously, I was responsible for vision and wholebody manipulation that powered the Atlas sequencing demos.

Publications

NeuralFeels with neural fields: Visuo-tactile perception for in-hand manipulation

NeuralFeels with neural fields: Visuo-tactile perception for in-hand manipulation

Sudharshan Suresh, Haozhi Qi, Tingfan Wu, Taosha Fan, Luis Pineda, Mike Lambeta, Jitendra Malik, Mrinal Kalakrishnan, Roberto Calandra, Michael Kaess, Joe Ortiz, Mustafa Mukadam

Science Robotics, 2024[Featured as journal cover]

CMU RoboticsAI at Meta

General In-Hand Object Rotation with Vision and Touch

Haozhi Qi, Brent Yi, Sudharshan Suresh, Mike Lambeta, Yi Ma, Roberto Calandra, Jitendra Malik

Proc. Conf. on Robot Learning, CoRL, 2023

New Scientist (UK)TU Dresden News

MidasTouch: Monte-Carlo inference over distributions across sliding touch

Sudharshan Suresh, Zilin Si, Stuart Anderson, Michael Kaess, Mustafa Mukadam

Proc. Conf. on Robot Learning, CoRL, 2022[Oral: 6% acceptance rate]

ShapeMap 3-D: Efficient shape mapping through dense touch and vision

Sudharshan Suresh, Zilin Si, Joshua Mangelson, Wenzhen Yuan, Michael Kaess

IEEE Intl. Conf. on Robotics and Automation, ICRA, 2022

Tactile SLAM: Real-time inference of shape and pose from planar pushing

Tactile SLAM: Real-time inference of shape and pose from planar pushing

Sudharshan Suresh, Maria Bauza, Peter Yu, Joshua Mangelson, Alberto Rodriguez, Michael Kaess

IEEE Intl. Conf. on Robotics and Automation, ICRA, 2021[Best paper award in service robotics finalist]

Active SLAM using 3D submap saliency for underwater volumetric exploration

Active SLAM using 3D submap saliency for underwater volumetric exploration

Sudharshan Suresh, Paloma Sodhi, Joshua Mangelson...

IEEE Intl. Conf. on Robotics and Automation, ICRA, 2020