I built Big Cola Bottle Inspection, a five-class detector for a simulated conveyor line, using a dataset I captured and annotated from scratch. Its validation reached 0.977 mAP50-95, and TensorRT FP16 reduced T4 GPU model-inference latency by 67% with no measurable mAP change.
For Chess Vision, I built a pipeline that converts chessboard screenshots to FEN using YOLOv8, OpenCV, and a classifier for all 64 squares. It achieved 99.3% exact-FEN accuracy on 150 unseen synthetic boards and shipped as a FastAPI web app in a CPU-only Docker container.
At Microsoft Student Club – Suez Canal University, I designed a six-month computer vision curriculum and mentored 28 students. I delivered four technical lectures and guided three hands-on projects covering CNNs, YOLO detection, transfer learning, augmentation, and model evaluation.
I'm studying Computer Science at Suez Canal University, with expected graduation in June 2027. My work spans computer vision model development, evaluation, and deployment with PyTorch, TensorRT, ONNX, and Docker.

