I am a software engineer and AI platform engineer based near Munich. At BayWa AG, I build agentic applications, process automation and the shared platform behind them. My background spans full-stack development, data systems and delivery across technical teams, users and business stakeholders.
I hold an M.Sc. in Applied Artificial Intelligence. My thesis investigated why computer-vision models fail to transfer reliably between agricultural datasets collected in different regions. I am particularly interested in intelligent systems that must remain dependable when their data and operating environment change.
- Relational representations for next-instrument anticipation - A compact public-data study comparing a temporal graph model with sequence baselines and topology controls on CAT-SG.
- Cross-dataset drift in soybean disease classification - Reproducible PyTorch experiments on cross-region transfer, calibration and failure analysis; the associated manuscript has been submitted for peer review.
- Computer vision for peatland monitoring and navigation - Segmentation, detection, depth estimation and remote-sensing workflows for ecological monitoring and outdoor navigation.
- Causal reinforcement learning for electric vehicles - Data and analysis for a systematic literature review, including an LLM-assisted Neo4j citation graph.
Python, PyTorch, OpenCV, scikit-learn | TypeScript, Java, Kotlin | Docker, Kubernetes, Azure | RAG, agentic workflows and process automation