Bio

Karl Tuyls is a Belgian computer scientist and senior researcher in multi-agent reinforcement learning and game theory. He is a co-founder of H, the Paris-headquartered agentic-AI company he established in 2023 with Charles Kantor, Laurent Sifre, Daan Wierstra, and Julien Perolat, and a former Research Director at Google DeepMind, where he led the Game Theory and Multi-Agent Team. As of May 2026, he is the director of Meta AI's Paris research lab, having joined Meta in late 2024 after his departure from H during the company's first-year founder-departure period.

At a glance

• Education: PhD in computer science, Vrije Universiteit Brussel (2004); postdoctoral position at Hasselt University.
• Current role: Director of Meta AI Paris research lab since late 2024 to early 2025; Professor of Computer Science at University of Liverpool (continuing); Guest Professor at KU Leuven (DTAI group).
• Past roles: Research Director, Google DeepMind (2017 to 2024); Co-founder, H (2023 to late 2024); Full Professor, University of Liverpool (2013 to 2017); Professor, Maastricht University (2005 to 2013).
• Key contributions: AAMAS 2023 opening keynote on "Multiagent learning: from fundamentals to foundation models"; multi-agent reinforcement-learning research and game-theoretic frameworks at DeepMind; co-founder of H.
• Personal site: karltuyls.net
• Google Scholar: Karl Tuyls
• LinkedIn: Karl Tuyls

Origins

Tuyls completed a PhD in computer science at Vrije Universiteit Brussel in 2004, after which he held a postdoctoral position at Hasselt University. The doctoral research was situated in the multi-agent reinforcement-learning area that became his principal lifelong research focus, with cross-disciplinary engagement with dynamic game theory and evolutionary methods for multi-agent learning.

He joined Maastricht University in 2005 as professor and built the multi-agent reinforcement-learning research program at the university through the late 2000s and early 2010s, establishing himself as a senior figure in the European multi-agent-systems research community.

Career

Tuyls moved to the University of Liverpool in 2013 as a full professor in the Department of Computer Science, where he led the multi-agent reinforcement-learning group through the mid-2010s. He continues to hold the professorship at Liverpool through 2026 alongside his subsequent industrial roles.

In 2017 he joined Google DeepMind as Research Director, leading the Game Theory and Multi-Agent Team across the lab's London and Paris offices. The DeepMind tenure produced extensive research on multi-agent reinforcement learning, game-theoretic foundations for cooperative and competitive agents, and the application of multi-agent methods to AlphaStar and other DeepMind research lines. He delivered the opening keynote at the AAMAS 2023 conference, "Multiagent learning: from fundamentals to foundation models," summarizing the DeepMind multi-agent research program in the context of the broader foundation-model transition.

In 2023 Tuyls left DeepMind to co-found H in Paris with Charles Kantor, Laurent Sifre, Daan Wierstra, and Julien Perolat. The five-co-founder configuration combined Kantor's academic-and-finance background with four senior DeepMind veterans, producing an unusually senior research-team composition for a European agentic-AI insurgent of the period. The May 2024 $220 million seed round, the largest AI seed in European history at the time, was led by Accel with Amazon, UiPath, Bpifrance, Innovation Endeavors, and individual investors including Bernard Arnault participating.

The founding period was disrupted by the departure of three of the five co-founders in late 2024. Tuyls, alongside Wierstra and Perolat, left H over what the company described as "operational differences." Sifted and other industry coverage detailed the departures as the principal first-year governance event for H. Tuyls subsequently took the role of director at Meta AI's Paris research lab. Per industry coverage of his transition, the appointment placed him at the senior research-direction layer at Meta's Paris office, where the FAIR organization has built a substantial European research presence alongside the broader Meta global research footprint.

The University of Liverpool professorship and the KU Leuven guest professorship continue alongside the Meta operating role, an academic-and-industrial dual track typical for senior multi-agent-systems researchers in the European research community.

Affiliations

• Vrije Universiteit Brussel: PhD candidate in computer science, awarded 2004.
• Hasselt University: Postdoctoral position, mid-2000s.
• Maastricht University: Professor, 2005 to 2013.
• University of Liverpool: Full Professor, 2013 to present.
• Google DeepMind: Research Director, Game Theory and Multi-Agent Team, 2017 to 2024.
• H: Co-founder, 2023 to late 2024.
• Meta AI Paris: Research lab director, late 2024 to early 2025 onward.
• KU Leuven (DTAI group): Guest Professor.

Notable contributions

• Multi-agent reinforcement-learning research at Google DeepMind (2017 to 2024). Led the Game Theory and Multi-Agent Team through the contemporary deep-reinforcement-learning consolidation. Research output included extensive multi-agent reinforcement-learning literature and game-theoretic frameworks for cooperative and competitive agents.
• AAMAS 2023 opening keynote, "Multiagent learning: from fundamentals to foundation models." The keynote at the principal autonomous-agents-and-multi-agent-systems conference summarized the DeepMind multi-agent research program in the foundation-model era.
• H co-founding (2023). Co-founded the Paris-headquartered agentic-AI company with Kantor, Sifre, Wierstra, and Perolat.
• Academic record. Senior multi-agent reinforcement-learning research record across the Maastricht, Liverpool, and DeepMind tenures, with the Google Scholar profile reflecting a substantial publication and citation record across game-theoretic learning, multi-agent systems, and reinforcement learning.
• Meta AI Paris director role. Direction of the Meta AI Paris research lab.

Position in the field

As of May 2026, Tuyls occupies a distinctive position bridging the academic multi-agent-systems community, the post-DeepMind senior-research diaspora, and the Meta AI Paris research environment. The combination of a continuing Liverpool professorship, the H founding tenure, and the Meta director role places him inside several of the principal European AI research network nodes simultaneously.

The Meta Paris director role positions Tuyls in the same Paris AI research environment that produced Mistral AI (founded by Meta FAIR Paris alumni Guillaume Lample and Timothée Lacroix with Arthur Mensch) and that hosts much of the European frontier-AI research workforce. Industry coverage has characterized the Meta Paris recruitment as part of the broader Meta strategy to anchor European AI research talent against the rising European insurgent labs and the Google DeepMind Paris office.

Outlook

Open questions over the next 6 to 18 months:

• Meta AI Paris research output. Continued publication and product output from the Meta AI Paris research lab under Tuyls's direction.
• Multi-agent reinforcement-learning research at Meta. Whether the Tuyls research program at Meta produces multi-agent-AI capability disclosures comparable to the Google DeepMind output of his Research Director tenure.
• Continued Liverpool and KU Leuven academic engagement. The pace of academic publication and research-student supervision alongside the Meta operating role.
• European AI research-talent dynamics. The competitive recruitment and retention dynamics among the European AI research labs, including Meta AI Paris, Google DeepMind Paris, Mistral AI, and the broader Paris AI research community.

Sources

• Karl Tuyls personal site. Personal academic site with current and past roles.
• Karl Tuyls on Google Scholar. Publication record and citation index.
• Karl Tuyls at Research.com. Computer Science research index entry covering academic posts and DeepMind tenure.
• Multiagent Learning Paradigms. Joint paper with Peter Stone on multi-agent learning paradigms.
• Sifted: Three cofounders leave French AI startup H. Coverage of the late-2024 H Company co-founder departures.
• TechCrunch: French AI startup H raises $220M seed round. Seed-round announcement covering the founding team.
• BNAIC/BeNeLearn 2022 invited speakers. Conference biographical entry covering academic and industrial roles.
• H (company) on Wikipedia. Company reference covering co-founder departures.