614 E. Daniel Street
Room 4131
Champaign, IL 61820
Medical AI for real-world care
Building medical AI for care as it actually unfolds.
I lead an interdisciplinary research group at UIUC that develops and evaluates AI for complex medical decision-making. We study how to benchmark what models truly understand, how they can reason across a patient’s evolving trajectory, and how realistic clinical simulations can expose failure modes before deployment.
My background in medicine, epidemiology, and health informatics shapes a translational approach: start with consequential clinical problems, build systems around the structure of care, and evaluate more than final-answer accuracy.
Assistant Professor, School of Information Sciences · Affiliate, Siebel School of Computing and Data Science · Affiliate, Health Care Engineering Systems Center
Previously: UW · AI2 · Google · Microsoft Research · Johns Hopkins
Three capabilities medical AI needs next.
Our projects connect rigorous evaluation, reasoning over time, and interactive environments into one research agenda.
Benchmark beyond accuracy
Develop evaluations that reveal whether models use the right evidence, recognize hidden concerns, and remain reliable under realistic constraints.
Evidence faithfulness · Long-form QA
See the program →Reason across patient journeys
Build systems that integrate longitudinal records, retain diagnostic experience, and revise decisions as evidence and patient state change.
See the program →Simulate the clinical world
Create interactive patients, care teams, and medical environments where actions change what happens next—so agents can be stress-tested safely.
Counterfactual agents · MedConceal
See the program →From research prototypes to useful medical AI.
We welcome collaborations with clinicians, health systems, researchers, and industry teams working on consequential problems in medicine.
Prospective students: I plan to recruit PhD and master’s students from the iSchool and CS programs beginning in Fall 2027. Please read the mentorship page and complete the linked interest form before reaching out about applications.
news
| Aug 24, 2026 | Two of our papers have been accepted to Findings of EMNLP 2026! Does Accuracy Equal Evidence? studies reasoning faithfulness under KV-cache compression, while LONGQAEVAL develops reliable, resource-conscious evaluation methods for long-form clinical QA. Congratulations to the students and collaborators who made this work possible! 🎉 |
|---|---|
| Jul 20, 2026 | A clean sweep at COLM 2026: 3/3 papers accepted! ReLay studies the benefits and risks of personalized health summaries; MedConceal tests how well models uncover and address patients’ hidden concerns; and SEA learns reusable diagnostic rules through jointly optimized reasoning and dual memory. I’ll be in San Francisco for the conference—please say hello! 🌉🎉 |
| May 26, 2026 | MedGuideX turns clinical practice guidelines into executable decision logic for LLM training, improving average clinical-reasoning accuracy by 10.28% and producing physician-preferred rationales. 🩺 |
| Apr 30, 2026 | Two papers are now published in the Journal of Biomedical Informatics: PlainQAFact introduces a sentence-aware, retrieval-augmented metric for checking factual consistency in biomedical plain-language summaries, while our crowdsourced study finds that LLM summaries can feel as clear as human-written ones yet produce worse actual comprehension. Reliable health communication must measure both factuality and what readers truly understand. 🧭 |
| Mar 29, 2026 | Our new counterfactual multi-agent framework tests how changing individual clinical findings shifts competing diagnoses, improving diagnostic accuracy and making agent reasoning more evidence-grounded on complex cases. 🔄 |