Projects

Trained binary classification models on a scientific image dataset of 10K samples to identify colon adenocarcinoma using PyTorch, iterating to find optimal model architectures and respective hyperparameters. Achieved diagnostic accuracy of 99%+ by leveraging the VGG16 & ResNet models with random forest. Deployed most accurate model to HuggingFace using Gradio for rapid inference & user-oriented design

Implemented a proxy for the Polygon REST API with AWS EC2, writing an HTTP server to handle requests using JavaScript, containerizing the NodeJS + Express app using Docker and setting up a reverse proxy using nginx. Currently developing React and NextJS front-end, providing user authentication & database integration using Supabase

Replicated the Airbus A320 plane using SolidWorks. Created exploded view diagrams with a BOM table. Also, everything was dimensioned to be model-sized because I plan on 3D printing it.

Designed a wearable biosensor for non-invasive intraocular pressure monitoring, securing $2000 in funding through the Ingenious+ & Toshiba ExploraVision Challenge, becoming top 24 of 2000+ teams across North America.