My Projects
A short list of what I've built at the intersection of deep learning, medical AI, and software engineering.
Featured
Frame to Flow
"Beyond the Last Frame: Temporal Deep Learning for Automated Grading of Retinal Inflammation in Fluorescein Angiography." A Vision Transformer (RETFound) turns each frame into an embedding, an unsupervised LDA step partitions each sequence into Early/Mid/Late phases, and a GRU reads the sequence in order to grade hyperfluorescence — beating a last-frame-only baseline on AUC across all six classification tasks.
BaseLine
A personal health tracking tool that analyzes lab results against your own baseline instead of population averages, using the Reference Change Value method laboratories have relied on for decades. Upload lab results via OCR, get an RCV-based read on what actually changed, all encrypted client-side before it ever reaches the server.
AI-Powered Diabetes Predictor
A web app that predicts diabetes risk from 8 health values — glucose level, BMI, age and more — and returns a result with a confidence score and charts explaining it. Built with a small team during the AI in Healthcare minor.