TrackForge
Particle track reconstruction with graph neural networks, framed as edge classification on detector hit graphs.
Hey, my name is
I study computer science at KIT. I am interested in physics, geometric deep learning and GNNs.

I started coding in 2016, at 12, with a summer course on Arduino microcontrollers at the Phaenovum Student Research Center in Lörrach. After that I moved on to HTML, CSS and JavaScript, as far as the time limit on our family computer allowed.
In 2019 my school decommissioned its old computers and I took one home. With a friend’s help I upgraded and overclocked it to modern standards (ish). I learned Python and Java on it, took Machine Learning courses at Phaenovum and worked on personal projects. I also took part in JugendForscht and the selection rounds for the Physics Olympiad.
In 2024 I started studying computer science, with physics as a minor, at the Karlsruhe Institute of Technology, supported by a scholarship from the Begabtenstiftung Informatik Karlsruhe.
During an internship at eschbach in 2025 I built an NLP proof of concept for search over industrial data. I went back in 2026, reworked it and brought it into production.
In 2025 I joined the perception team at KITcar.
During school I worked as a cave guide. I do track and field, and play guitar.
Particle track reconstruction with graph neural networks, framed as edge classification on detector hit graphs.
My final high school thesis, on handwritten digit recognition with very small datasets. Awarded for Mathematics and Computer Science at JugendForscht 2024.
A billiard game in the browser where ball collisions are randomised using a probability distribution.
A browser-based control panel for presenting PDF slides, with next-slide preview and notes.