I’m German and live near Munich. I got into ML through hands-on work in electrical systems, battery engineering, and embedded devices. Batteries are still what I know best, but I’m drawn to other physical systems where the data is messy, the constraints are real, and clean benchmarks do not tell the whole story.
Outside work, I have spent long stretches traveling across Asia and Australia.
Why physical systems
I got into batteries somewhat by accident on an 800 V/750 kW automotive prototype and stayed for the interplay of physics and hardware, tight safety margins, and the impact small improvements can have.
I see neural networks as generic meshes and backpropagation as a generic solver. ML earns its place when it shortens expensive engineering loops or captures what explicit models miss, but it must hold up over long horizons under realistic constraints.
How I like to work
I learn by building and going down the stack. I like taking accountability for a problem, making decisions, and having them challenged to find the best trade-off.
I prefer simple, inspectable systems over novelty for its own sake. An idea matters when it survives contact with data, hardware, and the people who use it.
Selected work
- batgrad: A template for training neural networks on battery data.
- Master’s thesis: Proposed neural model benchmark for battery state estimation
- 800 V automotive battery prototype: Owned electrical design through integration and successful validation
- Smart-glove prototype: Custom PCB and FPGA interface, paired with real-time ML