Youssef Donia

ML Engineer | Data Scientist with a CS degree and 4+ years shipping models that run in production — not just notebooks. Computer Vision, NLP, predictive maintenance, and embedded AI across semiconductor manufacturing, defense, transportation, and real estate. The thing about working in AI is that your job description will be different every year. I started when "doing AI" meant building neural networks from scratch. Then it was transformers, then fine-tuning LLMs, then chatbots and recommendation engines, then augmented reality (AR) damage trackers, and now it's agentic systems and MCPs. The stack keeps reshuffling and there is always something new to learn, so I never get bored. What that's looked like in practice: catching defects on semiconductor lines at Intel, transforming billions of rows of real estate data at Abodemine, building agentic AI compatible systems with real-time computer vision at Neology, and shipping production AI software for the Navy and Defense at Fathom5. The through-line across all of it is taking messy, high-stakes data and turning it into systems people can actually rely on under load. Outside of work, I'm a semi-pro pickleball player documenting my climb. I train the way I'd train a model: hypothesis, experiment, measure, iterate. It's taught me the same lesson that shows up everywhere else in my work: the fastest way to get better at anything is to be wrong on purpose, often, and quickly. I'll happily talk about that for longer than anyone wants me to. I grew up between Texas, Saudi Arabia, and Egypt, which is the short version of why I'm comfortable in cross-functional teams with people from very different backgrounds. It's also the longer version of why I'd rather work on things that improve real lives than chase a benchmark for its own sake. Ask me about coding on the peak of a mountain in Colorado, what it's like to be in the room with high-ranking Navy officials, or my current attempt to engineer a better backhand. Always open to meeting people who care about building real things, solving real problems, and learning out loud.