Selected for the NIH AIM-AHEAD Bridge2AI for Clinical Care Training Program (Cohort 2)—a competitive program that gives trainees hands-on experience working with real, multi-institutional patient data—I served as team lead for a project predicting how long premature infants stay in the NICU. Rather than chase accuracy with a black box, our team built a "glass-box" pipeline on a multi-site OMOP dataset, pairing an additive model (EBM) with explainable gradient boosting (XGBoost + SHAP) so clinicians can see which factors drive a baby's hospital stay. Both methods independently converged on the same clinically plausible drivers, and the interpretable approach surfaced data-quality issues early—pointing toward a transferable framework for trustworthy clinical AI in neonatal care.

Role: Team Lead · NIH AIM-AHEAD Bridge2AI for Clinical Care Training Program, Cohort 2

A digital health program bringing remote patient monitoring and care coordination to FQHCs serving rural and agricultural communities. Demonstrated improved diabetes and hypertension outcomes in underserved populations.

Role: Program Manager, CITRIS Health (2021–2024)

A mobile health application designed to support Certified Athletic Trainers in conducting concussion assessments. Extended evaluation capabilities beyond the sideline through remote consultation.

Award: Best Student Paper, AMIA Consumer Health WG (2017)

An open-source tool to help researchers convert existing biosketches to the NIH SciENcv format. Built to address a significant pain point where manual data entry was taking researchers 4+ hours.