My Projects & Technical Work

WasteWise – AI-Powered Waste Classification Web App

WasteWise is a full-stack web app I designed to classify waste items as Recyclable, Compostable, Trash, or Hazardous using AI-powered NLP models from Hugging Face. The app integrates barcode scanning and the OpenFoodFacts API for real-time product lookup and classification. I developed a mobile-responsive frontend using React.js and Tailwind CSS, while the backend is built with Node.js and Express. Firebase Authentication, Firestore logging, and interactive Recharts dashboards allow users to track their waste patterns over time.

Explore it on GitHub.

Chest X-ray Classification (Hackathon Project)

In this project, I built a multi-label image classification pipeline using pretrained CNNs (VGG16, ResNet50, EfficientNet) for chest X-ray datasets. I applied transfer learning, data augmentation, checkpointing, and early stopping to improve model generalization. A Flask API serves real-time predictions, while training metrics are visualized with Matplotlib and TensorBoard.

Project page on Devpost.

Urban Resilience.AI Lab – Machine Learning Research

As an undergraduate researcher, I developed machine learning models with TensorFlow and Scikit-learn to predict flood-induced pollutant dispersion from over 230 toxic facilities along the Texas coast. I analyzed 100,000+ geospatial and demographic data points using Pandas and NumPy to identify high-risk areas and underserved communities. I also designed spatial buffer zones (1–5 miles), integrated ML outputs with census data, and presented findings to stakeholders and policymakers.

This research enhanced my skills in geospatial analysis, predictive modeling, and translating technical results into actionable insights for real-world environmental issues. Learn more about the Urban Resilience.AI Lab.

Toyota Commercial Finance – AI Assistant Prototype

During my Software Engineering internship, I developed an AI assistant using large language models and NLP to automate internal support queries with 85% accuracy. I contributed to backend services using SQL and Microsoft Azure and optimized cloud-based data pipelines to enhance performance.

This project strengthened my experience in cloud computing, backend development, and applying AI to practical business problems.