Overview

I'm the Co-Founder and CEO of a stealth-stage machine learning infrastructure company.

Additionally, I co-organize a machine learning training program for engineers to learn about production-ready deep learning called Full Stack Deep Learning.

Previously, I was a researcher working at the intersection of machine learning and robotics. My research focused on applying deep reinforcement learning, generative models, and synthetic data to problems in robotic perception and control.

I did my PhD in Computer Science at UC Berkeley advised by Pieter Abbeel and was a research scientist at OpenAI for 3 years during my PhD. I have also been a management consultant at McKinsey and an Investment Partner at Dorm Room Fund.

Publications

• Real-World Robotic Perception and Control Using Synthetic Data,

  Josh Tobin. PhD Dissertation, UC Berkeley, Computer Science, May 2019.

• Geometry-Aware Neural Rendering,

  Josh Tobin, OpenAI Robotics, Pieter Abbeel. 33rd Conference on Neural Information Processing Systems (NeurIPS 2019), Vancouver, Canada, December 2019. (Oral). [slides]

• Domain Randomization and Generative Models for Robotic Grasping,

  Josh Tobin, Lukas Biewald, Rocky Duan, Marcin Andrychowicz, Ankur Handa, Vikash Kumar, Bob McGrew, Alex Ray, Jonas Schneider, Peter Welinder, Wojciech Zaremba, Pieter Abbeel. In the proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), Madrid, Spain, October 2018. [pdf, video]

• Learning Dexterous In-Hand Manipulation,

  OpenAI: Marcin Andrychowicz, Bowen Baker, Maciek Chociej, Rafal Jozefowicz, Bob McGrew, Jakub Pachocki, Arthur Petron, Matthias Plappert, Glenn Powell, Alex Ray, Jonas Schneider, Szymon Sidor, Josh Tobin, Peter Welinder, Lilian Weng, Wojciech Zaremba. arXiv:1808.00177, August 2018. [pdf, blog post, video 1, video 2]

• Multi-Goal Reinforcement Learning: Challenging Robotics Environments and Request for Research,

  Matthias Plappert, Marcin Andrychowicz, Alex Ray, Bob McGrew, Bowen Baker, Glenn Powell, Jonas Schneider, Josh Tobin, Maciek Chociej, Peter Welinder, Vikash Kumar, Wojciech Zaremba. arXiv:1802.09464, February 2018. [pdf, blog post]

• Hindsight Experience Replay,

  Marcin Andrychowicz, Filip Wolski, Alex Ray, Jonas Schneider, Rachel Fong, Peter Welinder, Bob McGrew, Josh Tobin, Pieter Abbeel, Wojciech Zaremba. In Neural Information Processing Systems (NIPS), Long Beach, CA, December 2017. [pdf, video]

• Domain Randomization for Transferring Deep Neural Networks from Simulation to the Real World,

  Josh Tobin, Rachel Fong, Alex Ray, Jonas Schneider, Wojciech Zaremba, Pieter Abbeel. In the proceedings of the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), Vancouver, Canada, October 2017. [pdf, video 1, video 2, video 3, Blog post 1, Blog post 2, Blog post 3]

• Transfer from Simulation to Real World through Learning Deep Inverse Dynamics Model,

  Paul Christiano, Zain Shah, Igor Mordatch, Jonas Schneider, Trevor Blackwell, Josh Tobin, Pieter Abbeel, Wojciech Zaremba. arXiv:1610.03518, October 2016. [pdf]

Slides & Guides

• A Missing Link in the ML Infrastructure Stack

  Where is ML going as an industry? Is infrastructure keeping up? And what tools might still be missing? From Stanford MLSys Seminar, February 2021.

• Geometry-Aware Neural Rendering

  Slides from my talk at NeurIPS 2019.

• Randomization and the Reality Gap

  A guide to transfering robotic policies from simulation to the real world, with a particular focus on domain randomization. From a talk in Berkeley's CS287, November 2019.

• Beyond Domain Randomization

  How is domain randomization being used now, and where should we go from here? Slides from a talk at the Sim2Real Workshop at RSS, June 2019.

• Troubleshooting Deep Neural Networks

  A decision tree for building bug-free models and improving their performance. Originally from a talk at Full Stack Deep Learning, March 2019.