Julian Hoßbach
Ph.D. student at the Institute for Computational Physics, University of Stuttgart
I work on developing machine learning methods in the field of nanopore-based peptide sensing, which aims to identify individual peptides in a solution. More concretely, I develop methods to detect very subtle temporal features in the recordings from these experiments, which are noisy and vary in length. Picking up these features reliably would allow recognizing single peptides directly in a sample, which could make it possible to detect cancer biomarkers earlier and more cheaply. My publications are listed on Google Scholar.
I supervise Bachelor’s and Master’s theses and run the exercises for the lectures Physik auf dem Computer, Computergrundlagen and Simulation Methods in Physics.
At the Institute for Computational Physics, I also co-administrate and supervise our local HPC cluster.
I like to support open source projects where possible. I maintain python-lsp-ruff, a Ruff plugin for the Python language server, and a few packages on the AUR. I have also contributed to ESPResSo, the soft matter simulation package developed at our institute.
Background
- 2017 – 2021
- B.Sc. Physics, University of Stuttgart
- 2021 – 2024
- M.Sc. Physics, University of Stuttgart
- since 2025
- Ph.D. student at the Institute for Computational Physics, University of Stuttgart
Current topics of interest
Things I would like to understand better, and what I do in my spare time.
Physics
- Nonlinear optics
- Electrodynamics in the microscopic regime
- Meteorology
Machine learning
- Reinforcement Learning
- Diffusion models
- Embedded ML
Programming
- Python, C and C++, which I have worked a lot with over the years
- Rust, which I am tinkering with lately, mostly on embedded targets.
Hobby projects
- ESP32 powered smart home gadgets
- Playing with ultra low-power electronics, such as e-ink displays
Music
- Playing the piano, the French horn and the baritone horn
Contact
You can reach me at julian.hossbach@gmx.de. My PGP key is available on keys.openpgp.org and has the fingerprint
0A9A CC5D DBEB 9FE5 5F28 7F35 89BC 4C56 7C0F 982E