At Amazon, I architected and deployed a hazardous materials inventory compliance system that blocked non-compliant storage operations and unlocked approximately $100 million in annual operational savings. I also developed real-time inventory pipelines and automated warehouse workflows.
As a Software Engineering Intern at Amazon, I developed explainability methods and visualization tools for an NLP transformer model using SHAP values. At the University of Connecticut, I implemented a CNN-based traffic anomaly detection model that improved performance by 4.5% over traditional baseline methods.
