Sara’s research lies at the intersection of multimodal and self-supervised learning, with the goal of learning meaningful representations from limited and heterogeneous medical data. She earned her PhD in AI for Medical Imaging from the University of Toronto, where her research focused on generating generalizable and explainable representations of complex clinical data through multimodal and self-supervised learning. During her PhD, she developed contrastive learning frameworks that integrate brain MRI with radiology reports, leveraging complementary visual and textual information to improve downstream diagnostic performance and visual explainability. Her research also addressed key challenges in clinical settings, including limited labeled data, false negatives in contrastive learning, and domain shift. Her research has been recognized through several awards and fellowships, including the Ontario Graduate Scholarship.