Youssef M.

OCR & Document AI Engineer | Computer Vision | Production ML I build production-ready OCR, Document AI, and Computer Vision systems that turn images and documents into reliable, structured data and usable AI features. I work end-to-end — from data preparation and model development to API integration and deployment. What I can build: • OCR and document extraction pipelines • PDF/image → structured JSON • Invoice, receipt, and document processing • Image classification and object detection • Image captioning and vision-based AI • Medical imaging / segmentation pipelines • VLM/LLM-powered document understanding • FastAPI + Docker ML APIs • Model evaluation, optimization, and production deployment My typical workflow: Data & preprocessing → model selection/training → evaluation → API → Docker/deployment Tech stack: Python, PyTorch, TensorFlow, Hugging Face, OpenCV, FastAPI, Docker, MLflow, Transformers, U-Net / nnU-Net Recent work includes: • Built and deployed an image captioning system through FastAPI • Developed an AI-powered PDF/document extraction system • Improved an existing CNN classification model • Contributed to a nnU-Net medical imaging pipeline I focus on building systems that are not only accurate in a notebook, but practical to integrate into a real product. Have a project in mind? Send me: 1. A brief description of the problem 2. A sample document/image or dataset 3. The target outcome (prototype, production API, model improvement, or evaluation) I'll assess the approach, identify the main technical challenges, and suggest the most practical implementation.