AI Engineer | Deep Learning & Computer Vision Specialist | Egypt
I'm a results-driven AI engineer and computer vision specialist, passionate about deploying advanced deep learning solutions on edge devices such as the NVIDIA Jetson Xavier NX. My journey at Assiut University culminated in a graduation project focused on real-world automotive safety, contributing to Egypt’s Vision 2030 for sustainable innovation. I excel in model optimization, sensor fusion, scalable deployment, and bridging the gap between research and impactful real-world applications.
- AI Deployment: NVIDIA Jetson Xavier NX, TensorRT, quantization
- Deep Learning: CNNs, transfer learning, model optimization
- Computer Vision: Object detection, pose estimation, multi-sensor fusion
- Data Science: scikit-learn, Optuna, advanced visualization
- Software Engineering: C++, Python, Qt, responsive GUIs, speech recognition
|
Integrated Driver Road Monitoring System (HeadPose & EyeGaze) Real-time driver attention monitoring system for automotive safety, leveraging MobileNetV2 and TensorRT for 31 FPS inference on Jetson Xavier NX. Features advanced deep learning, spatial attention, and deployment for Euro NCAP 2025 standards. Tech: PyTorch, NVIDIA Jetson, TensorRT, OpenCV |
|
Brain Tumor Classification: VGG16 & Augmented MRI Automated brain tumor detection using VGG16 and data augmentation on MRI images. Achieved 98.34% accuracy by fine-tuning select layers. Tech: TensorFlow, Keras, Medical Imaging |
|
Exploring Regression Models with UMAP and Ridge Applied UMAP and Ridge regression to predict house prices, optimizing RMSE to 0.22 and demonstrating advanced data science and visualization. Tech: Python, scikit-learn, UMAP |
|
CNN: Early Stopping, Data Augmentation, and Visualizations Trained a CNN for digit recognition (28,000 images), achieving 99.1% accuracy and global top-400 ranking, with advanced training visualizations. Tech: Keras, Data Augmentation, Visualization |
|
Optimizing XGBoost for Fraud Detection with Optuna Built a high-performance credit card fraud detection system on 218,000 user records, optimized with Optuna for an AUC score of 0.83878. Tech: Python, XGBoost, Optuna |
|
AI Task Manager Cross-platform AI-powered task scheduler with C++ backbone, Python AI (Random Forest, 95% accuracy), Qt GUI, and voice input. Tech: C++, Python, Qt, Speech Recognition |