About me.
I’m Jens — here, I share work and notes on machine learning, applied AI systems, LLM engineering, and building reliable AI systems.
I hold two master’s degrees — one in mathematics and one in computer science — and my curiosity has led me into adjacent fields like psychology. At Heidelberg University, I researched machine learning under distribution shift: “How can we make reliable predictions when circumstances change?” (robustness, domain generalization) and “How can we detect when shifts and failures occur?” (OOD detection, failure detection, selective classification). I published on these topics and completed my PhD in the process.
Currently, I work as a Senior Data Scientist at Diamant Software, focusing on Natural Language Processing (NLP) and applied AI. Recent industry projects include an autonomous AI assistant, RAG architectures, evaluation pipelines, and agent systems.