Projects

A selection of research, engineering, and regulatory-science work spanning clinical AI, pharmacovigilance, and biomedical data systems.

Featured

Flagship Projects

🚀 End-to-End

Cough Acoustics

ML-based respiratory screening from audio signals

Built with

Python Signal Processing ML Pipelines

DigitalTwin: Cough-Based Respiratory Disease Detection

Machine learning platform for respiratory screening using cough acoustics.

END-TO-END TEAM

GitHub

🚀 End-to-End

20M+

FAERS adverse-event records analyzed for signal detection

Signal methods

ROR PRR BCPNN

GLP1-Sentinel: Pharmacovigilance Signal Detection

End-to-end pharmacovigilance pipeline with regulatory-grade signal detection across FAERS.

END-TO-END SOLO

GitHub

🧠 ML

🏆 2nd Prize

Hackathon award for CNN-based binding-affinity prediction

Built with

CNN PyTorch Drug Discovery

MaxBindAI

CNN-based ligand affinity prediction for drug discovery — 2nd Prize Winner.

ML FORK/COLLAB

GitHub

Regulatory

Regulatory Sciences & Intelligence

⚖️ Regulatory Sciences

FDA Pathway Intelligence

for Digital Health & Devices

FDA Device / Digital Health Pathway Intelligence

One of the projects closest to my heart: a regulatory-science decision-support dashboard that maps likely FDA pathways, identifies predicate devices, and surfaces evidence gaps for early-stage digital health and medical device strategy.

REGULATORY SCIENCE SOLO

GitHub

Publications

Published Research

📖 In preparation

Multi-omics Kidney Transplant Cohort

Led statistical design and end-to-end analysis of a multimodal kidney-transplant longitudinal cohort integrating 42-marker CyTOF (~32.3M cells), metabolomics, and clinical data. Identified candidate immune-metabolic biomarkers for pre-transplant allograft rejection risk stratification, and built reproducible R pipelines for publication-ready analyses in collaboration with transplant clinicians.

IN PREPARATION

📖 In preparation

Scleroderma scRNA-seq — CellChat Analysis

Analyzed single-cell RNA-seq data from scleroderma-associated interstitial lung disease and used CellChat to identify disease-specific cell–cell signaling networks.

IN PREPARATION

Modeling