At Coca-Cola Egypt, I contributed machine learning work to an AI-based analytics project that improved fraud detection and reduced fraudulent activity by 5%. I also built and trained an image classification model for product categorization.
For my Smart School System graduation project, I designed and implemented an end-to-end AI system for student monitoring and safety. It included face recognition with a Siamese network, liveness detection, violence detection, and uniform compliance classification.
I built an Agentic NL2SQL System that turns natural-language questions into executable SQL queries across a multi-table database. The workflow includes schema routing, SQL generation and validation, self-correction, and natural-language answers.
My other projects include Arabic sentiment analysis, news group classification, and movie popularity prediction. I’ve also developed a full-stack web application for Modern Broadcast Center using Next.js, Prisma, and PostgreSQL.

