PeopleofSuperintelligence

Researcher

Michał Bortkiewicz

Also known as Michal Bortkiewicz

Michał Bortkiewicz is a reinforcement learning researcher who works on goal-conditioned and self-supervised RL. As first-listed author, he released JaxGCRL in 2024, a GPU-accelerated benchmark and codebase that trains agents for millions of environment steps in minutes on a single GPU and appeared as an ICLR 2025 spotlight. In 2025 he coauthored "1000 Layer Networks for Self-Supervised RL" with Benjamin Eysenbach and others, which showed that depth up to 1024 layers can significantly boost performance and won a Best Paper Award at NeurIPS 2025. He and Tomasz Trzciński received the Breakthrough Research Award named for Stanisław Ulam in March 2026. He is a PhD student at Warsaw University of Technology and a visiting researcher at Princeton University.

Timeline

  1. 2022

    Emergency termination in hierarchical RL

    In 2022, he is the first-listed author of "Emergency action termination for immediate reaction in hierarchical reinforcement learning," written with Jakub Łyskawa, Paweł Wawrzyński, Mateusz Ostaszewski, Artur Grudkowski and Tomasz Trzciński. The method verifies at the higher policy level whether a subgoal is still valid and replaces it when it becomes inadequate, and the paper tests it on seven benchmark environments.

  2. 2024

    Two ICML 2024 papers

    He coauthors two papers at ICML 2024: "Overestimation, Overfitting, and Plasticity in Actor-Critic: the Bitter Lesson of Reinforcement Learning" with Michal Nauman, Piotr Miłoś, Tomasz Trzciński and others, and the spotlight "Fine-Tuning Reinforcement Learning Models is Secretly a Forgetting Mitigation Problem" with Maciej Wołczyk, Bartłomiej Cupiał and others.

  3. 2024–2025

    The JaxGCRL benchmark

    He is the first-listed author of "Accelerating Goal-Conditioned RL Algorithms and Research," posted in August 2024 with Władysław Pałucki, Vivek Myers, Tadeusz Dziarmaga, Tomasz Arczewski, Łukasz Kuciński and Benjamin Eysenbach. The paper releases JaxGCRL, a codebase and benchmark that trains agents for millions of environment steps in minutes on a single GPU and cuts training time by up to 22×, and it appears as a spotlight at ICLR 2025.

  4. June 2025

    Visiting researcher at Princeton

    In June 2025, he starts a stay at Princeton University as a Visiting Student Research Collaborator. He works closely with Benjamin Eysenbach there and continues his PhD at Warsaw University of Technology, where Tomasz Trzciński and Piotr Miłoś advise him.

  5. October 2025

    A study of experience stitching

    In October 2025, he posts "Is Temporal Difference Learning the Gold Standard for Stitching in RL?" with Władysław Pałucki, Mateusz Ostaszewski and Benjamin Eysenbach; he and Pałucki contribute equally. The paper shows that Monte Carlo methods can also achieve experience stitching, and that the gap with temporal difference methods is smaller than the gap between small and large neural networks.

  6. November 2025

    A NeurIPS 2025 Best Paper

    In November 2025, "1000 Layer Networks for Self-Supervised RL: Scaling Depth Can Enable New Goal-Reaching Capabilities," which he coauthors with Kevin Wang, Ishaan Javali, Tomasz Trzciński and Benjamin Eysenbach, receives a Best Paper Award at NeurIPS 2025. The paper shows that increasing network depth up to 1024 layers can significantly boost performance in unsupervised goal-conditioned reinforcement learning, and Trzciński credits Bortkiewicz's contribution and the earlier JaxGCRL benchmark.

  7. March 2026

    The Ulam Breakthrough Research Award

    In March 2026, he and Prof. Tomasz Trzciński receive the Breakthrough Research Award named for Stanisław Ulam. The prize is presented at the first Ulam Awards gala, held during the MIT Universal AI Summit in Warsaw.

  8. As of October 2026

    PhD student and Princeton visitor

    He is a PhD student at Warsaw University of Technology and a visiting researcher at Princeton University, working on unsupervised and goal-conditioned reinforcement learning, representation learning for efficient exploration, and methods to mitigate catastrophic forgetting and improve plasticity. In 2026, he coauthors "On the Role of Computation in Reinforcement Learning" with Raj Ghugare, Alicja Ziarko and Benjamin Eysenbach, a paper for ICML 2026.

Career

  1. Since 2025Princeton UniversityVisiting Student Research Collaborator

Education

  1. Warsaw University of TechnologyPhD

From the biography.

  1. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other universityIncludes Princeton University. The Massachusetts Institute of Technology trained or employed 44 of the 187 AI major-prize winners, more than any other university. We compare doctorate and 5+ year employment links, career timing and current workplaces for 187 major-prize and 1,111 best-paper winners.
  2. 84% of people shaping AI with recorded undergraduate and doctoral schools changed institutions for their doctorateIncludes Princeton University. 84% of people shaping AI with recorded undergraduate and doctoral schools changed institutions for their doctorate. Across 2,116 people with both an undergraduate degree and a doctorate recorded, we trace which undergraduate schools feed which doctoral programs, how often students stay at the same school, and how paths differ by birth-country region.
  3. More people shaping AI studied at Massachusetts Institute of Technology than at any other institutionIncludes Princeton University. We estimate where the people shaping AI studied and where their careers led. Massachusetts Institute of Technology leads with 467 former students, including 291 recorded doctorates.
  4. The median h-index is 48 at universities and 14 at frontier AI labsIncludes Princeton University. The median h-index is 48 at universities and 14 at frontier AI labs. We compare publishing careers, recent citations, dual appointments and job moves across 2,420 university researchers and 308 frontier-lab researchers.

Colleagues

People named in this biography, and people with dated positions at the same organization in the same years.

Sources and credits

Photo: We have no published photo of Michał Bortkiewicz yet. Send us a free-licensed photo.

  1. [2211.06351] Emergency action termination for immediate reaction in hierarchical reinforcement learning arxiv.org. Checked 21 October 2026.
  2. Michal Bortkiewicz - Michal Bortkiewicz michalbortkiewicz.github.io. Checked 21 October 2026.
  3. [2408.11052] Accelerating Goal-Conditioned RL Algorithms and Research arxiv.org. Checked 21 October 2026.
  4. [2510.21995] Is Temporal Difference Learning the Gold Standard for Stitching in RL? arxiv.org. Checked 21 October 2026.
  5. Announcing the NeurIPS 2025 Best Paper Awards – NeurIPS Blog blog.neurips.cc. Checked 21 October 2026.
  6. NeurIPS 2025 Awards neurips.cc. Checked 21 October 2026.
  7. [2503.14858] 1000 Layer Networks for Self-Supervised RL: Scaling Depth Can Enable New Goal-Reaching Capabilities arxiv.org. Checked 21 October 2026.
  8. NeurIPS 2025 Best Paper Awards in the hands of our researchers / News / Home - Warsaw University of Technology eng.pw.edu.pl. Checked 21 October 2026.
  9. Prof. Tomasz Trzciński i jego doktorant, mgr inż Michał Bortkiewicz zdobywcami nagrody Ulam Award / Aktualności / Strona główna - Instytut Informatyki Wydziału Elektroniki i Technik Informacyjnych ii.pw.edu.pl. Checked 21 October 2026.
  10. [2602.05999] On the Role of Computation in Reinforcement Learning arxiv.org. Checked 21 October 2026.