Avi Amin — Portfolio (CS, Math & Engineering)

CNN image classifier for identifying diseases in apple crops. Trained on a plant disease dataset; explores applied machine learning for precision agriculture.

Hurricane building-damage classifier from before/after satellite imagery, with Grad-CAM explainability. Siamese ResNet18 trained on real xBD disaster data.

Monocular 3D speed and acceleration tracker — tracks any object in a video and recovers its real-world motion from a single camera, physics-validated against gravity.

Real-time wildfire and smoke detection with a fine-tuned YOLOv8 model, built for aerial drone footage.

ML pipeline predicting the 2026 FIFA World Cup — a from-scratch ELO rating engine feeding an XGBoost classifier, run through 1,000 Monte Carlo tournament simulations.