Cataract Detection System

Most medical imaging tools are black boxes you can't touch. I wanted to build something end-to-end — from raw eye images to a deployable API that actually works. Used YOLOv8 for detection, PyTorch for training, and wrapped the whole thing in a Flask API. The model hit 98.1% detection accuracy. The bigger win was understanding what it takes to go from a Jupyter notebook to something you could actually hand to a doctor.