Tuomas Haarnoja
Co-Founder at Eka Robotics
Cambridge, Massachusetts, United States
Summary
👤
Senior🎓
Top SchoolTuomas Haarnoja is a robotics researcher and entrepreneur with 11+ years of experience building full-stack machine learning control systems, from low-level electromechanical control to high-level behavior planning. A UC Berkeley PhD advised by Pieter Abbeel and Sergey Levine, he co-created the widely adopted Soft Actor-Critic algorithm that has become a staple in robot learning libraries and attracted mainstream coverage. He has held research roles at DeepMind and Google Brain, applying sample-efficient RL to real-world locomotion and manipulation, and now co-founds Eka Robotics to bring those advances into products. Tuomas’s background in classical control and hands-on electromechanical design gives him a rare ability to translate learning algorithms into robust hardware implementations. His SAC repo contributions show practical attention to reproducibility and performance (replay buffer refactors, deterministic test sampling), reflecting an engineer who balances theory, code hygiene, and deployment realities.
11 years of coding experience
11 years of employment as a software developer
Doctor of Philosophy - PhD Computer Science, Doctor of Philosophy - PhD Computer Science at University of California, BerkeleyMS Space Science and Technology, MS Space Science and Technology at Aalto UniversityBachelor of Science - BS Automation and Systems Technology, Bachelor of Science - BS Automation and Systems Technology at Teknillinen korkeakoulu-Tekniska högskolan
English, Finnish
Github Skills (8)
machine-learning10
trainings10
python10
reinforcement-learning10
modeling10
tensorflow9
hyperparameter-tuning9
algorithms8
Programming languages (4)
Github contributions (5)
haarnoja/sac
Nov 2017 - Dec 2018
Soft Actor-Critic
Role in this project:
ML Engineer
ML EngineerContributions:72 commits, 9 PRs, 13 pushes in 1 year 1 month
Contributions summary:Tuomas primarily contributed to the development and refinement of a Soft Actor-Critic (SAC) reinforcement learning algorithm, as indicated by the code modifications. Their work involved fixing author information, cleaning up and adapting example scripts, and removing redundant hyperparameters. They also refactored the replay buffer implementation, and added deterministic sampling at test time, indicating an understanding of RL algorithm implementation, experimental setup, and performance optimization.
soft-actor-criticactor-criticactorcritic
haarnoja/softqlearning
Jun 2017 - Jan 2019
Reinforcement Learning with Deep Energy-Based Policies
Contributions:45 commits, 8 PRs, 11 pushes in 1 year 7 months
reinforcement-learningdeep-reinforcement-learningpoliciesreinforcementenergy
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.