Karim Elgazar

Built and deployed deep learning models for precise medical image segmentation (lung cancer, brain tumors, abdominal organs), providing critical support for diagnostics.

Developed and optimized chest X-ray disease classification models, significantly enhancing diagnostic accuracy and efficiency for healthcare professionals.

Deployed AI models as robust APIs with LitServe, ensuring seamless front-end integration and real-time inference for clinical applications.

Designed and implemented an LLM-powered system to suggest personalized treatments based on comprehensive disease detection and patient assessments.

Trained and fine-tuned large-scale models using Azure ML and AWS SageMaker, optimizing performance and scalability for diverse medical datasets.

Utilized AWS S3 for comprehensive data and model versioning, ensuring integrity, collaboration, and auditability throughout the AI development lifecycle.

Completed over 117 projects, accumulating 858+ hours and generating $50,000+ in verified revenue for diverse international clients.

Developed and deployed bespoke computer vision solutions for clients in geospatial analysis, medical imaging, and document intelligence, ensuring high-quality results and client satisfaction.

Demonstrated advanced expertise in Python, OpenCV, TensorFlow, and PyTorch to deliver innovative, production-ready AI applications within tight deadlines.

Consistently achieved high client satisfaction ratings, earning and maintaining Top-Rated Plus freelancer status through exceptional project management and value delivery.

Developed and optimized deep learning models for object detection and segmentation (trees, pools, cars, waste) from satellite imagery, significantly improving geospatial analysis efficiency.

Implemented robust object tracking algorithms (centroid tracking, BYTE-SORT) and distance estimation techniques, enhancing precision for large-scale aerial data analysis.

Curated a large-scale dataset of millions of images using open-source resources, augmentation, and human-in-the-loop annotation, ensuring high-quality data for model training.

Deployed trained models as scalable API endpoints on GCP and AWS SageMaker, enabling real-time data processing and seamless integration with client systems.

Managed and secured extensive datasets and models on AWS S3, implementing version control and ensuring data integrity for critical infrastructure projects.