Bio

I am a second-year Computer Science DPhil student at the University of Oxford, supervised by Sam Staton. My research interests lie (mostly) at the intersection of mathematics and theoretical computer science - such as the semantics of programming languages, Markov chain Monte Carlo methods, reinforcement learning theory and information geometry. I currently split my time between doing research on categorical semantics of probabilistic programming languages and working on AI safety.

I am generously funded by Open Phillantropy's Early-Career Funding for Improving the Long-Term Future.

Before coming to Oxford, I studied Mathematics (BSc, MSc) and Computer Science (BSc) at the University of Warsaw, worked in a hedge fund in London, tried studying Philosophy at the University of Warsaw (but dropped out pretty quickly), Medicine at the Medical University of Lodz (and dropped out even quicker), developed NLP models in 2016s before Transformers had taken over, tried my luck at mechanism design during the blockchain frenzy of 2017s, stationed at the polar circle while adding a few additional bugs to Office365, and in general engaged in BFS over DFS while I could.

Publications

• Limitations of Agents Simulated by Predictive Models, Raymond Douglas, Jacek Karwowski, Chan Bae, Andis Draguns, Victoria Krakovna (to appear at the ICLR 2024 Workshop on LLM Agents)
• Goodhart's Law in Reinforcement Learning, Jacek Karwowski, Oliver Hayman, Xingjian Bai, Klaus Kiendlhofer, Charlie Griffin, Joar Skalse (to appear at ICLR 2024)
• Probabilistic programming interfaces for random graphs: Markov categories, graphons, and nominal sets, Nate Ackerman, Cameron Freer, Younesse Kaddar, Jacek Karwowski, Sean Moss, Daniel Roy, Sam Staton, Hongseok Yang (published at POPL 2024)
• Symmetries in reversible programming: from symmetric rig groupoids to reversible programming languages, Vikraman Choudhury, Jacek Karwowski, Amr Sabry (published at POPL 2022)