Scholar
Katarzyna Kobalczyk
Google Scholar ID: PyIbICgAAAAJ
University of Cambridge
Citations & Impact
Citations
37
H-index
3
i10-index
2
Publications
12
Co-authors
4
list available
Publications
Resume (English only)
Academic Achievements
- 2025: ICML 2025, 'Preference Learning for AI Alignment: a Causal Perspective'
- 2025: ICML 2025, 'The Synergy of LLMs & RL Unlocks Offline Learning of Generalizable Language-Conditioned Policies with Low-fidelity Data'
- 2025: ICLR 2025, 'Active Task Disambiguation with LLMs'
- 2025: ICLR 2025, 'Towards Automated Knowledge Integration From Human-Interpretable Representations'
- 2022: Institute of Mathematics and its Applications (IMA) Prize, University of Warwick
- 2022: Warwick Statistics Prize, University of Warwick
- 2021: Outstanding Academic Excellence Prize, University of Warwick
Research Experience
- June 2025 – September 2025: Research Scientist Intern, Meta
- June 2023 – September 2023: Quantitative Research Intern, Citadel
- June 2022 – September 2022: Data Science Research Intern, G-Research
- August 2021 – October 2021: URSS Undergraduate Researcher, University of Warwick
- June 2021 – August 2021: Data Insights Unit Intern, Schroders
- June 2021 – September 2021: Data Science Intern, Shell
Education
- 2023-2027 (expected): PhD in Applied Mathematics and Theoretical Physics, University of Cambridge
- 2022-2023: MASt in Mathematical Statistics, University of Cambridge
- 2019-2022: BSc in Mathematics and Statistics, University of Warwick
Background
- PhD candidate in the Machine Learning and AI lab at the University of Cambridge, Department of Applied Mathematics and Theoretical Physics. Her research aims to make AI systems more data-efficient, adaptive, and aligned with humans. She studies how models, especially large language models, can represent human knowledge and make decisions consistent with both empirical data and contextual cues. Her work combines ideas from meta-learning, probabilistic ML, and Bayesian experimental design to build learning frameworks that provide principled uncertainty estimates and support more trustworthy decision-making.
Miscellany
- She is actively engaged in a collaborative partnership with Eedi, working alongside industry experts to enhance the effectiveness of studying and teaching mathematics among school-age children.
