Brief CV

I'm a PhD student with the German Research Center for Artificial Intelligence (DFKI) in Darmstadt.

I want robots with near-human dexterity in everyday household tasks. To achieve this, I'm making flexible, compliant robots that can adapt to unforeseen situations, and designing algorithms to teach them humanlike manipulation skills.

Email me at aditya@bhatts.org.

Education

• University of Freiburg / MSc / Computer Science
  2016 - 2019
• DA-IICT / BTech / Information and Comuniction Technology
  2008 - 2012

Papers & Articles

• Surprisingly Robust In-Hand Manipulation: An Empirical Study
  Aditya Bhatt*, Adrian Sieler*, Steffen Puhlmann, Oliver Brock (Jul. 2021, RSS 2021)
• CrossNorm: Normalization for Off-Policy TD Reinforcement Learning 
  Aditya Bhatt, Max Argus, Artemij Amiranashvili, and Thomas Brox (Feb. 2019)
• Artificial Intelligence for Prosthetics - Challenge Solutions (NeurIPS 2018 competition book)  Kidziński et al. (Invited to present at the NeurIPS 2018 workshop)
• Stabilizing Off-Policy TD with Feature Normalization
  Aditya Bhatt (Master Thesis, Apr 2019)

Talks

• How the Body Shapes the Way we Manipulate (Embodied Intelligence 2021 workshop)
• Surprisingly Robust In-Hand Manipulation: An Empirical Study (Robotics: Science and Systems 2021 spotlight, jointly with Adrian Sieler)

Skills

• Solid real-world Python, Javascript, C++ programming skills. Strong software engineering experience.
• Very comfortable with PyTorch. Familiar with TensorFlow.
• Training Deep Networks and RL Agents
• Designing Robotic Manipulation Skills
• Building Pneumatic and Tendon-Driven Compliant Robots