Engineering, at its best, is an act of translation — turning messy physical reality into systems that behave with intention. I work at that boundary: close enough to hardware to understand its constraints, fluent enough in ML to know what the data can actually tell you.
I am a mechatronics student at THWS Würzburg-Schweinfurt, a hackathon finalist, and a builder who is most alive when the problem resists clean decomposition. I ship fast, think in feedback loops, and have never met a stack I wasn't willing to learn.
Local reverse proxy that sits transparently between any LLM client and the upstream API — intercepts calls, tracks token usage per session, and auto-compacts context at configurable soft/hard thresholds before limits are hit.
6-stage vision pipeline automating dataset generation for 16 blood tube classes — RealSense depth camera, MobileSAM segmentation, and Pydantic-versioned reproducible exports.
ESP32
Arduino
MATLAB
Simulink
OpenCV
MobileSAM
YOLOv8
PyTorch
Scikit-learn
Python
C/C++
Node.js
TypeScript