Salmon Angelo

Through my academic projects, I have gained hands-on experience in machine learning, computer vision, backend integration, and real-time AI systems. I have worked extensively with Python, Pandas, Scikit-learn, FastAPI, OpenCV, and machine learning models such as Random Forest and XGBoost. One of my major projects, FraudSentinel AI, involved building a hybrid fraud detection system using ensemble learning, RAG pipelines, and local LLM integration. I also developed CrowdCare, a real-time crowd risk monitoring system using YOLOv8 and OpenCV, where I optimized inference latency for CPU-based deployment. These projects strengthened my understanding of data preprocessing, model evaluation, explainability, and scalable AI workflows.