关于
I’m a Principal Researcher at Microsoft Research, working broadly on AI for health, with a focus on multimodal AI and foundation models.
I have led research projects and collaborations with healthcare and AI organizations including Providence Health, Epic Systems, and OpenAI, spanning large language models, clinical data, radiology and pathology.
Selected work includes:
- Vision & multi-modal foundation models for biomedical data: GigaPath (Nature), GigaTIME (Cell), BiomedParse (Nature Methods), LLaVA-Med (NeurIPS), BioViL (ECCV), and BiomedCLIP (NEJM AI).
- Joint research with Epic Systems on large-scale patient foundation model trained on 100M+ patient records.
- Joint research with OpenAI on healthcare applications of frontier language models.
- Large language models for health: Medprompt o1/GPT-4 with OpenAI, PubMedBERT (Best paper, ACM Health), CXR-BERT (ECCV), and OncoBERT (Cell Patterns).
- Clinical AI and agentic systems with healthcare partners including Providence and Stanford: Research on AI systems for real-world clinical workflows, including molecular tumor board decision support, clinical trial matching, clinical guideline reasoning, and medical literature synthesis, spanning patient data, multimodal models, LLMs, and agentic systems.