All Projects

Finance Anomaly Detector

An end-to-end fraud detection system built with machine learning. Features a React frontend for real-time transaction analysis, FastAPI backend for high-performance API serving, and XGBoost model achieving 92.2% ROC-AUC accuracy. Includes Docker deployment, comprehensive testing, and production-ready architecture. Processes transactions in under 200ms with confidence scoring and risk level classification.

Ant Colony Simulation

Multi-agent simulation demonstrating emergent intelligence and collective problem-solving. Built with JavaScript and D3.js for real-time visualization. Simulates 1,000+ concurrent agents following simple rules that lead to complex behaviors. Explores concepts like swarm intelligence, pathfinding algorithms, and emergent patterns in distributed systems.

MovieLens Analysis

Comprehensive data analysis and machine learning project on the MovieLens dataset. Includes exploratory data analysis, recommendation system implementation, and predictive modeling. Uses collaborative filtering, matrix factorization, and deep learning approaches. Features data visualization, user behavior analysis, and model performance comparison across different algorithms.