Mahmoud E.

Are you looking to transform your business with cutting-edge artificial intelligence? I'm a Senior AI Engineer at Elsewedy Electric, one of the region's largest energy and infrastructure groups, where I architect enterprise-grade AI solutions for large-scale industrial operations. I bring that same enterprise rigor to freelance work, specializing in Generative AI, Natural Language Processing (NLP), and cloud-based deployments. My expertise lies in building scalable Generative AI applications. I seamlessly integrate and orchestrate Azure OpenAI models using advanced prompt engineering and RAG (Retrieval-Augmented Generation) architectures to solve complex, business-specific use cases. What I deliver: Cost-saving automation: I engineered a centralized Azure Speech Service module adopted company-wide. This initiative drove a 50% reduction in operational costs. End-to-end model development: I design and train AI algorithms for tasks such as Natural Language Processing (NLP) and Computer Vision. I also fine-tune pre-trained models for specific use cases. Seamless cloud deployment: I deploy and maintain robust AI pipelines on Microsoft Azure. This includes leveraging Azure Speech Services for real-time transcription. Robust data infrastructure: I have implemented data and ETL pipelines. This enabled the development of cloud-based data warehouses for real-time analytics. Featured Project — AI-Powered Event Photo Matcher: Developed a facial recognition pipeline, deployed on a scalable AWS instance, that automatically filters and distributes event photos. Attendees simply upload a profile picture, and the system scans bulk event uploads to instantly deliver their personalized photo gallery. This project involved heavy use of Python and Computer Vision libraries like OpenCV, driving an increase in face-matching accuracy from 60% to 95%. The matching process is exceptionally fast because it leverages extracting and storing lightweight face embeddings for each picture, allowing the system to rapidly compare dense vectors instead of processing the entire photo during the search phase, which improved processing performance by 10x compared to the original method. Technical Arsenal: AI, ML & Deep Learning: Python, TensorFlow, PyTorch, Scikit-learn, OpenCV Generative AI: Azure OpenAI, LLMs, Transformers, RAG architectures Cloud & Big Data: Microsoft Azure, Google Cloud, AWS, SQL, Hadoop, Spark