Faisal Mahmood

Faisal Mahmood, PhD is Associate Professor of Pathology at Harvard Medical School and Director of the Computational Pathology and AI Laboratory at Brigham and Women's Hospital (BWH/HMS). He is one of the world's leading researchers in computational pathology and AI-driven oncology, best known for developing large-scale foundation models that enable AI systems to interpret pathology images with human-expert-level proficiency across diverse cancer types. His laboratory developed CONCH (CONtrastive learning from Captions for Histopathology), a vision-language foundation model for pathology pre-trained on over 1.17 million image-caption pairs, which achieved state-of-the-art performance in cancer subtype classification, survival prediction, and zero-shot pathology tasks. He subsequently introduced TITAN, a multimodal whole-slide foundation model that integrates slide-level and patch-level visual representations with pathology-specific language grounding, enabling generalist pathology AI with strong transfer learning capabilities. Dr. Mahmood pioneered weakly supervised multiple instance learning approaches for slide-level prediction from gigapixel pathology images, enabling cancer diagnosis, molecular subtype classification, and prognosis prediction without pixel-level annotation. His group has demonstrated AI models that can predict microsatellite instability (MSI), BRCA mutation status, genomic signatures, and immunotherapy response directly from H&E-stained tissue sections. He has published extensively in Nature Medicine, Nature Biomedical Engineering, Cancer Cell, and NeurIPS, and his models are among the most widely adopted open-source tools in computational pathology.