AI People Wiki

Builders & open source

Tom Schaul

Senior staff research scientist; leads a reinforcement learning team, Google DeepMind

  • Senior staff research scientist; leads a reinforcement learning team, Google DeepMind13

Tom Schaul is a senior staff research scientist at Google DeepMind and leads a reinforcement learning team there, having joined DeepMind in 2013 while it was a start-up. Schaul's research covers exploration, hierarchy, credit assignment and learning dynamics in deep reinforcement learning, along with black-box optimization. Schaul is first author of "Prioritized Experience Replay" (2015), which replays important transitions more often than uniform sampling would, and of "The Phenomenon of Policy Churn" (NeurIPS 2022).123

At a glance

Current role
Senior staff research scientist; leads a reinforcement learning team
Sector
Company
Education
NYU
Work listed
2
Sources cited
5
Profile reviewed

Career

Schaul completed a PhD from TU Munich in 2011 under the supervision of Jürgen Schmidhuber at the Swiss AI Lab IDSIA. From 2011 to 2013 Schaul was a postdoc in Yann LeCun's lab at New York University's Courant Institute, and in 2013 joined DeepMind while it was a start-up. Schaul is now a senior staff research scientist who heads a reinforcement learning team at Google DeepMind.1

Schaul's service to the field includes serving as program chair for the Reinforcement Learning Conference (RLC) 2026, keynote chair and senior area chair for RLC 2025, and area chair for ICML and NeurIPS 2024. Schaul has also co-organized a Dagstuhl workshop on AI for the social good and taught at the International Summer School on Artificial Intelligence and Games.1

Research

Schaul's research covers many aspects of deep reinforcement learning, including exploration, hierarchy, goal discovery, learning dynamics, generalization and credit assignment, as well as black-box optimization and minimal hyperparameter tuning, with games as a favorite application area. Papers include Universal Value Function Approximators (ICML 2015), Prioritized Experience Replay (ICLR 2016), "Ray Interference" (RLDM 2019) on plateaus in deep RL, and "The Phenomenon of Policy Churn" (NeurIPS 2022).124

Schaul also co-authored DeepMind's 2019 Nature paper on reaching grandmaster level in StarCraft II with multi-agent reinforcement learning, a 2020 Nature Communications perspective on AI for social good, a 2024 ICML position paper on open-endedness presented as an oral, and the Gemini 2.5 technical report in 2025.1

Work and contributions

  • Prioritized Experience Replay Published research

    First author, with John Quan, Ioannis Antonoglou and David Silver, of the 2015 paper developing a framework for prioritizing experience in replay memory.2

  • PyBrain Open-source code

    Contributed recurrent neural network, reinforcement learning and black-box optimization code to the PyBrain Python machine learning library, through which Schaul releases many research algorithms and benchmarks.5

Timeline

  1. Completed a PhD at TU Munich under Jürgen Schmidhuber at IDSIA, then began a postdoc at NYU's Courant Institute.1
  2. Joined DeepMind while it was a start-up.1
  3. Presented "Universal Value Function Approximators" at ICML 2015.1
  4. Posted "Prioritized Experience Replay" as first author.2
  5. Co-authored the Nature paper on grandmaster-level StarCraft II play with multi-agent reinforcement learning.1
  6. Posted "The Phenomenon of Policy Churn" as first author.4
  7. Presented a position paper on open-endedness as an oral at ICML 2024.1
  8. Began serving as program chair for the Reinforcement Learning Conference 2026.1

Other documented roles

Roles the sources document that are not confirmed as current.

Where they work

Suggest a correction or removal

Profiles cover public professional work only, and each statement cites a source. If something is out of date, missing or wrong, tell us and link a source. People can also ask for their profile to be corrected or removed.

References

5 sources · profile reviewed Oct 21, 2026 · data updated Oct 21, 2026

  1. Tom Schaul, PhD — Google DeepMind (homepage)Tom Schaul · Official page · accessed Oct 21, 2026
  2. Prioritized Experience Replay (arXiv 1511.05952)arXiv · Paper · published Nov 18, 2015 · accessed Oct 21, 2026
  3. schaul (Tom Schaul) on GitHubGitHub · Official page · accessed Oct 21, 2026
  4. The Phenomenon of Policy Churn (arXiv 2206.00730)arXiv · Paper · published Jun 1, 2022 · accessed Oct 21, 2026
  5. Software — Tom SchaulTom Schaul · Official page · accessed Oct 21, 2026