Biography
I lead the machine learning platform team at Trade Republic, where we run financial crime detection models in real time inside the transaction path of one of Europe’s largest neobrokers. I got here by way of a PhD on interpretable machine learning at Bielefeld University and years of production ML engineering at super.ai, working on LLM systems and the cloud infrastructure underneath them. What drives me is making systems run efficiently: I like taking something that already works and making it faster and cheaper, whether that means shaving inference latency or cutting infrastructure cost. Away from the keyboard I lift heavy things, run half marathons, walk up mountains, and lose evenings to strategy games.
- Advanced Machine Learning Techniques
- Cutting-edge Software Engineering Practices
- Machine Learning Operations (MLOps)
- Innovative Cloud Application Development
Dr. rer. nat. Machine Learning (magna cum laude), 2021
Bielefeld University
B.Sc & M.Sc. in Bioinformatics and Genome Research, 2014 & 2016
Bielefeld University
Recent Posts
Experience
Tech Lead of the Machine Learning Platform team supporting the Financial Crime (FinCrime) element, serving Trade Republic’s customer base at scale (promoted from Senior MLE in Feb 2026).
- Operate a real-time ML platform running many models continuously inside the transaction path, built for low latency and high availability
- Enabled engineers and data scientists across the company to ship models self-serve, guarded by canary rollouts, automated health checks, and progressive traffic shifting
- Drove migration to Kubernetes and built a ClickHouse-based real-time feature store
- Technical Leadership: Guided the team’s technical direction, fostered a culture of agility and continuous improvement, and mentored engineers.
- Document Intelligence & LLMs: Applied state-of-the-art large language models for text matching and information extraction across millions of records, and built techniques for accurate extraction from diverse document types.
- Platform & Cloud: Designed a scalable, event-driven AI model backend service; led migration of a heterogeneous legacy AWS stack to Kubernetes with Terraform (IaC), improving fault tolerance and establishing SOC2 compliance; developed and extended open-source Python SDKs for the platform.
Doctoral research on interpretable machine learning (feature relevance bounds), with peer-reviewed publications in Neurocomputing and at ESANN and IEEE CIBCB. Included a six-month research stay in Prof. Ester’s data mining group at Simon Fraser University, Vancouver.
- Developed and released FRI, an open-source Python library for interpretable feature selection on high-dimensional biomedical data
- Modelling, analysis, development, and deployment; scientific writing and presentation; teaching
