I've built zero-shot computer vision pipelines for weed detection with the University of Hohenheim, using CLIP, Grounding DINO, and SAM 2 to reduce manual annotation in precision agriculture.
At Fraunhofer IIS, I automated training and evaluation workflows for AI-based audio noise reduction, building Python and Bash data pipelines, QA scripts, and reproducible research workflows. I also developed web interfaces for cloud-based audio encoding prototypes using React, TypeScript, JavaScript, and Webpack.
I'm completing an M.Sc. in Data Science at Friedrich-Alexander-Universität Erlangen-Nürnberg, where my thesis evaluates zero-shot vision models through large-scale GPU experiments and deployment-focused analysis.

