Biography

I am excited by the promise of artificial intelligence to help advance science and medicine, from improving cancer diagnosis to accelerating drug discovery. Since 2020, I’ve been working on foundation models for pathology, which became the basis of several state-of-the-art cancer detection, subtyping, and digital biomarker tools. I am particularly interested in models that combine multi-modal data, e.g. images and text, to learn abstract, unified, and interpretable representations of the world and perform complex reasoning with them. This capability is paramount for the successful application of AI in medicine since such systems will have to integrate diverse, complex, and connected medical data with specialised and structured medical knowledge. In my research at Cambridge, I aim to understand the fundamental principles and limitations of learning from multi-modal data for complex reasoning, investigate how models can learn interpretable and useful representations without explicit supervision, and transfer my findings to real-world applications, such as biomedical image analysis.

Previous Education

Warsaw University of Technology Electrical Engineering
Swiss Federal Institute of Technology Electrical Engineering
National Technical University Power Engineering