Real-time fraud detection system using ensemble ML models, featuring streaming data processing, explainable AI with SHAP, and production-ready deployment with FastAPI and Docker.

Real-time UPI fraud detection system (0.8953 ROC-AUC) with <500ms FastAPI scoring, 480+ temporal features, and budget-aware alerts under fintech constraints

Real-time AI-enabled hand gesture recognition system for deaf and mute communication using OpenCV, MediaPipe, and Python

Código fuente: Análisis de Vuelos basado en trabajo de Valliappa Lakshmanan.

Streamlit app for credit card fraud detection using XGBoost.

Streaming Evaluation Protocol (Prequential + Resource Footprint + Calibration Drift)

🛡️ Detect and prevent fraud in real-time with an ensemble system, achieving high accuracy and significant ROI for enhanced security solutions.