Abbas Mustafa

• • Developed and optimized deep learning algorithms for automated medical condition detection using 2D and 3D CT scan analysis.

• • Implemented volumetric segmentation pipelines using nnU-Net and custom Transformer models, processing med ical data across Sagittal and Coronal axes to improve diagnostic accuracy.

• • Leveraged MONAI for Multi-Planar Reconstruction (MPR) and data augmentation, resulting in a 25% reduction in training time via GPU kernel optimization.

• • Engineered containerized microservices for clinical deployment, ensuring HIPAA-compliant inference and 100% reproducibility.

• • Architected and deployed a mission-critical logistics and immigration management platform for 10,000+ inter national travelers at a high-sensitivity International Border Crossing.

• • Engineered a high-performance FastAPI backend and Flutter mobile application for real-time ID card scanning, biometric verification, and automated data retrieval for secure presence tracking.

• • Deployed and maintained the system on a localized **off-grid server environment, managing a high-concurrency network of 40+ devices under peak load with zero downtime.

• • Designed and deployed a Computer Vision pipeline for Metropolitan Surveillance Infrastructure, focusing on real-time traffic violation detection across a centralized surveillance grid.

• • Developed an Automatic Number Plate Recognition (ANPR) engine and Red Light Violation Detection (RLVD) module integrated with an automated electronic ticketing system.

• • Optimized YOLO-based object detection models for high-concurrency video streams on industrial edge hardware.