I am a mathematician and programmer at heart. My passion is to develop high performing algorithms and software for challenging problems.

During my Ph.D in applied Mathematics at the Industrially focused Mathematical Modelling CDT (InFoMM) at the University of Oxford.
My work focused on applied novel reinforcement learning methods alongside classical collocation methods to complex, and long horizon, optimal control problems.
In optimal control, we are interested in interacting with complex systems which accept inputs from either a human or computer. Most optimal control problems of practical interest are too complex to be solved analytically. My work is focused on designing reinforcement learning environments for complex and long-time-horizon problems. The reinforcement learning algorithms can then be used in conjunction with classical collocation methods to design novel high performing algorithms.

I studied applied mathematics at TU-Dortmund, University of Cambridge, University of Oxford.

One of my biggest side-projects is to explore the application of reinforcement learning methods applied to trading financial instruments, such as cryptographic currencies. I've designed several strategies, from classical ones such as dual momentum, or turtle trading to complex reinforcement learning based algorithms using several time frames as data sources.

I have also co-authored a paper on the maxnodf algorithm, which approximates the NP-hard problem of nestedness maximisation in ecological networks using simulated annealing algorithms. My main contribution to this algorithm was the discovery of large a performance increase by using incremental updates to improve the time complexity of this algorithm form O(N^4) to O(N^2). The results can be viewed in this paper. Several earlier results are included in this paper.