Hi! I am a medical AI researcher at NIH, supported by the K99/R00 Pathway to Independence Award ($997,000, one of about three awarded per year by NLM). My long-term goal is to democratize medical knowledge by combining rigorous evaluation, evidence-grounded retrieval and generation, and AI agents to provide accurate and verifiable information to clinicians, patients, and scientists.

My first-author research (e.g., PubMedQA and TrialGPT) has appeared in Nature Methods, Nature Communications, Nature Protocols, npj Digital Medicine, NeurIPS, ACL, EMNLP, and SIGIR. In total, our work has over 8,000 citations on Google Scholar, 65%+ from the papers that I led or co-led.

We developed the MedCPT foundation models that have been downloaded over 8 million times & serving millions of users through NLM products, and PubTator Database that has been accessed over 1 billion times. Our work has also been adopted by Google, Microsoft & OpenAI, Meta, Anthropic, Nvidia; and featured by NIH News Releases & Catalyst, POLITICO, Nature, Nature Biotech, MedScape.

Our research was highlighted by NIH in its budget justification sent to the United States Congress.