Developed an AI-based fraud detection system using Python and Scikit-learn to identify potentially fraudulent financial transactions. Built a machine learning pipeline for data preparation, model training, prediction, and fraud analysis, with a Streamlit-based interface for practical interaction and monitoring.

Built a production-oriented multi-modal lost-and-found platform that intelligently matches lost and found items using image, text, location, category, and time signals. Implemented AI-assisted matching, geospatial indexing, secure evidence verification, cryptographic handover codes, privacy controls, automated security testing, mobile camera/GPS integration, and CI/CD workflows. The project demonstrates practical experience in AI integration, full-stack development, application security, scalable architecture, and real-world problem solving.