I build decision-critical AI/ML systems — the kind where the model's output gates whether software ships, whether a vehicle is safe to deploy, or whether a supply chain holds.
Now
I lead ML and data science for virtual testing at Latitude AI, Ford's autonomous vehicle subsidiary. My team builds simulation and performance-prediction systems that gate every AV software release. Before Latitude, I worked on reinforcement learning for industrial automation at Microsoft, supply chain forecasting at Amazon, and production planning optimization at AMD.
Writing & Teaching
I wrote Mastering Reinforcement Learning with Python (Packt, 2020), an Amazon bestseller covering deep RL through real-world applications. I've taught statistics and operations research at UT Austin and Texas State, and I hold a Ph.D. and M.S. from Boston University and a B.S. from Bilkent University (full scholarship).
What I'm thinking about
We're at the beginning of a shift where AI stops being a tool you use and starts becoming infrastructure that adapts to you. Software that reshapes itself to how you work. Education that meets each learner where they are. Research that accelerates not by replacing scientists but by removing the friction between questions and answers. I think the most interesting work in AI right now isn't making models bigger — it's making systems that personalize intelligently, at every layer, for every user. That's the thread connecting my day job in AV simulation, rlbook.ai, and what I'll write about next.