Research & Projects

End-to-end pipeline for cleaning, modeling, and visualizing clinical datasets with reproducible workflows using MIMIC-IV.

Created a local web app to stay up to date on publications within subject areas and from specific researchers.

Developed a regime-aware quantitative portfolio management application using statistical jump models, XBoost regime predictors, and Sharpe ratio optimization.

Combined decision-focused learning with rolling horizon optimization to create a CT scanning regimen planning simulator.

Mapped the Princeton ORFE Department's co-authorship network using a three-layer BFS web scraping. Researchers were organized by their subject areas and application domains of interest using zero-shot text classification.