Mohammad Atiullah Shaikh

Smart EcoSort: Waste Segregation System | Python, Computer Vision, IoT Aug 2024 – Oct 2024 • Developed an AI-powered waste segregation system, improving classification efficiency through accurate sorting of biodegradable and non-biodegradable waste • Integrated a YOLO-based machine learning model, achieving 92% accuracy for real-time waste identification in an IoT-enabled dual-compartment dustbin • Designed a user-friendly interface, increasing system adaptability by 50% • Conducted 150+ tests to refine accuracy, addressing image quality and sensor precision challenges Oil Spill Detection Using Deep Learning Models | Python, YOLO, Roboflow Jun 2024 – Aug 2024 • Developed a deep learning-based oil spill detection system with an F1 score of 0.89 using SAR satellite imagery • Preprocessed and annotated 10,000+ images via Roboflow, enhancing training data quality • Trained YOLO v8, YOLO v9, and UNET, improving detection accuracy by 10% over baseline models • Optimized models, reducing false positives by 20% and boosting segmentation accuracy with DeepLabV3 • Analyzed 50+ training iterations with confusion matrices, achieving a 15% efficiency gain