Overview
My research investigates domain generalisation in computational pathology, with the long-term goal of building multimodal AI systems that can robustly integrate histopathology, molecular, and clinical data across healthcare institutions. My current work focuses on understanding how domain shift is encoded within pathology foundation model representations and how this knowledge can be used to design simpler and more robust machine learning methods. By combining representation theory with multimodal learning, I aim to develop clinically reliable models that remain effective under real-world distribution shifts.