Yu Gu

顾禹 (Gù Yǔ) — Aiden

I’m a researcher and entrepreneur focused on turning foundational AI models into real-world solutions across science, technology, and industry. Now at Microsoft Research and Microsoft Health & LIfe Science, I previously co-founded a Series C AI startup, scaling foundational research into enterprise-grade solutions.

My work centers on large-scale foundation models , multimodal learning , robotics and agentic AI frameworks , with real-world deployments across healthcare, life sciences, and beyond.

Selected Highlights

• PubMedBERT : Pioneered one of the first domain-adaptive language models in biomedicine ( ACM HLTH Best Paper ), with 20M+ downloads , 2000+ citations , and integration into Azure Text Analytics for Health, serving 10+ major institutions.
• BiomedParse : Created a best-in-class universal segmentation model ( Nature Methods ), delivering state-of-the-art performance across CT, MRI, and pathology. 1M+ monthlydownloads ; #1 in CVPR 2025 3D Biomedical Segmentation Challenge .
• BiomedJourney : Developed a generative model for counterfactual image generation, enabling robust interpretation of disease progression and deployment in data-scarce settings.
• Multi-Agent Systems : Built agentic LLM frameworks for collaborative clinical reasoning, showcased at the World Economic Forum and Microsoft Build , improving tumor board decision-making in complex oncology cases.
• Startup : Co-founded and scaled an AI healthcare company to enterprise adoption, culminating in a Series C acquisition .

Published in Nature , Cell , ICLR , ACL , and more. Reviewer for Nature journals, TMI, NeurIPS , etc. Work featured by Forbes, CNBC, and the World Economic Forum .