Building in Health AI at Duke.
Interested in the intersection of medicine and machine learning.
A CNN to diagnose eye diseases from retinal images with 89% accuracy. Currently in aquisition process.
ArchiveA universal explainability layer for "black box" clinical AI models. Architected a two-part XAI system using Grad-CAM and fine-tuning MedGemma for NLP interpretation.
ArchiveA desktop-native AI agent that reads your screen, predicts your next action, and executes in real-time. Now open source.
Open SourceA modular measuring device using optical rotary encoders and microcontrollers. Awarded Best Presentation at Duke EGR101 Symposium; slated for $10M Duke accessibility project.
Archive