Earlier recognition of depression
Using patient-reported signals, large language models, and longitudinal data to identify emerging depression risk and evaluate clinically responsible surveillance strategies.
Cardiometabolic care
Clinical informatics × data science × patient experience
Patient signals.
Better decisions.
Developing patient-centered AI that models evolving health and behavior from multimodal patient-reported signals to improve clinical decision-making.
Patient-reported signals
Multimodal data
Longitudinal modeling
Clinical decision support
Research vision
Care informed by lived experience.
I study what patients report between clinical encounters and how those signals can help us recognize changing health earlier.
My research develops patient-centered AI that transforms symptoms, concerns, treatment experiences, and behaviors into actionable information for earlier detection and personalized care. I combine large language models, natural language processing, and longitudinal machine learning across patient messages, clinical conversations, audio, and emerging data sources.
The goal is not simply to predict an outcome. It is to understand how health evolves over time, communicate risk meaningfully, and support decisions that reflect each patient’s needs, preferences, and daily experience.
01 / IDENTIFYPatient prioritiesCapture symptoms, concerns, barriers, and treatment experiences.
02 / DISCOVERChanging riskModel persistence, co-occurrence, escalation, and trajectories.
03 / SUPPORTClinical decisionsTranslate evidence into understandable, useful action.
Research portfolio
From observation to action.
A connected program spanning concern identification, longitudinal surveillance, early detection, and individualized decision support.
Mental health
Cardiometabolic care
Stroke signals and patient action
Characterizing symptom timing and escalation before stroke, then translating risk patterns into understandable patient-facing guidance.
Risk translation
Evolving symptoms before colorectal cancer
Studying how symptoms persist, co-occur, and transition before diagnosis to inform more personalized early-detection strategies.
Early detection
Dynamic models for personalized decisions
Integrating patient-reported, clinical, and multimodal data to represent changing health states and compare alternative screening, monitoring, and treatment strategies.
Future care
Publications & scholarship
Selected work.
Highlighted research advancing patient-centered AI, clinical communication, and responsible decision support.
Kim J, Chen ML, Rezaei SJ, et al.
Kim J, Ma SP, Chen ML, et al.
Kim J, Chen ML, Rezaei SJ, et al.
Kim J, Leonte KG, Chen ML, et al.
Kim J, Chen ML, Rezaei SJ, et al.
Prospective students & researchers
Build patient-centered AI with us.
I welcome trainees who are curious about how patient experience, clinical data, and computational methods can come together to improve care.
Student researchers
PhD, master’s, and undergraduate students can contribute to NLP, longitudinal modeling, multimodal AI, clinical informatics, data analysis, and research communication.
Postdoctoral research fellows
Lead interdisciplinary studies, develop independent directions, and translate computational advances into meaningful, patient-centered research.
News
Updates.
Research-group news, collaborations, presentations, publications, and recruiting.
Joining UCSF. Beginning a new position as Assistant Professor in the Division of Clinical Informatics and Digital Transformation at the UCSF School of Medicine.
Microsoft–Stanford DAX Copilot Research Award. Awarded $100,000 to advance research using clinical audio and transcript data.
Stanford Cancer Institute Innovation Award. Awarded $50,000 for “Closing gaps in cancer care disparities by listening to patients’ own voices.”
AMIA Amplify Scientific Program Committee. Selected to serve on the Scientific Program Committee for two consecutive years, 2026 and 2027.
AMIA NLP Working Group Leadership Team. Serving on the leadership team to advance collaboration and scholarship in clinical natural language processing.
NIH/NIMH K01 Career Development Award. “Leveraging Large Language Models and Machine Learning Algorithms to Assess Depression and Anxiety Symptoms and Risks for Patients with Cardiovascular Disease or Diabetes Mellitus.”
Interested in research or collaboration?
Share your background, interests, and what you hope to explore.