I have completed a PhD at Delft University of Technology, where I develop model-based reinforcement learning algorithms that combine parametric MPC policies with end-to-end learning for decision-making under uncertainty.
My work spans transportation, energy systems, multi-agent systems, and nonmyopic global optimisation. I've developed probabilistically safe learning methods using Gaussian processes and control barrier functions, and co-authored six journal and four conference publications.
I maintain open-source projects including mpcrl and csnlp, contribute to leap-c and acados, and have also built embedded control, computer vision, and race-critical data software for rehabilitation robotics and Ferrari F1 wind-tunnel testing.
