AI for Science & Society

Ai for digital pathology

Foundation models and interpretable tissue phenotyping for large-scale cancer imaging data.

At TU Delft, the University of Edinburgh, the Institute of Cancer Research, and the Royal Marsden Hospital, I have contributed to computational pathology programmes spanning approximately 200 TB of imaging data.

Projects include rhabdomyosarcoma tissue phenotyping, CD31 vessel segmentation, contrastive representation learning, and the development of foundation-model approaches for clinically meaningful imaging biomarkers.