George Drakoulas, PhD
I sit between two worlds that rarely share a language: physics-based simulation and production machine learning. My work is making them talk to each other.
I did a PhD in mechanical engineering and aeronautics — the slow, careful business of simulating how physical structures behave, where I built machine-learning surrogate models that ran up to 100× faster than the simulations they stood in for. Alongside and since, I've spent seven years putting machine learning into production for heavy industry, where a wrong answer has a cost measured in steel and schedule, not just a dashboard metric.
That combination is unusual. Most people in AI have never validated a finite element model; most people who have, don't ship ML. Sitting in both is what lets me build surrogate models an engineer will actually trust, and know where they stop being trustworthy.
Today I work in maritime R&D at Damen, on the modelling and machine learning side — surrogate models for hydromechanics, predictive maintenance across sensor fleets, and tools that compress engineering work that used to take days.
Freelance and advisory
I take on a small number of freelance and advisory engagements alongside my role, where the problem sits at the edge of simulation and ML. If that sounds like what you're stuck on, get in touch.
Background
- PhD, Mechanical Engineering & Aeronautics — University of Patras (2021–2025)
- Integrated MEng, Mechanical Engineering — University of Patras, graduated top 2% of class
- Seven years of production ML across maritime, automotive and manufacturing — Damen, Toyota Motor Europe, and my own venture
- Full CV (PDF)