Johannes Wilhelm

I’m a Computer Science student at RWTH Aachen University, currently completing my Master’s. I’m interested in Machine Learning, Quantitative Finance, and Neuroscience, especially enjoying problems that require both creative and technical problem-solving.

Parallel to my studies, I have worked at Forschungszentrum Jülich as a Student Research Assistant, improving the optimization framework L2L for large-scale simulations of spiking neural networks. Before that, I worked at Postera Capital as a Quantitative Research Analyst, where I developed predictive pricing models for cryptocurrency assets. I built an ML pipeline for time-series forecasting, which later became a key part of my work there.

For my Master’s thesis at RWTH, I worked with Prof. Naumann and Neil Kichler on Efficient Higher-Order Sobolev Training. I proposed and evaluated scalable ways to incorporate curvature and higher-order information into neural network surrogate training. For my Bachelor’s thesis, I worked with Prof. Borchers and Dr. Hueber on User-Perspective Rendering for Augmented Reality, where I explored how face-tracking can be used to enhance immersion in AR applications through novel rendering techniques.

On this blog, you’ll find posts about some of my projects and thoughts on ML, algorithmic trading, and whatever else I find interesting. The goal is to document my work, organize my thoughts, and share ideas.