Model
Make assumptions and conservation laws explicit.
Teaching & supervision
Teaching mathematics, numerical methods, and data-driven engineering.
Students learn to translate a physical question into equations and data, choose a method, implement it reproducibly, test it against evidence, and explain when the result should not be trusted.
Data
Understand origin, uncertainty, bias, and information value.
Code
Build results that can be inspected, tested, and reproduced.
Judgment
Connect error and validity to the engineering decision.
Teaching portfolio
From mathematical foundations to open computational practice.
Short conceptual introductions lead into guided calculation, open-ended projects, and explicit reflection on assumptions and uncertainty.
Mathematics II for Electrical Engineering
Lead lecturer at TU Darmstadt for a large foundational course, coordinating cumulative content and exercises for students with varied preparation.
Machine Learning in Fluid Mechanics
Co-conception, development, and organization of a new specialized master’s course for ten students, connecting field data, supervised learning, reduced representations, neural networks, and model assessment.
LEIA Methods I and II
Independently conceived and delivered lectures and exercises on Lagrangian/Eulerian interface-advection methods, implemented algorithms, and verification cases.
Research software, data & OpenFOAM–ML
Seminars, datathons, and recurring hackathons connecting version control, continuous integration, metadata, HPC, CFD, and machine learning.
Agentic computational workflows
Develop, test, document, and audit numerical methods with agent systems and LLMs while V&V, provenance, reproducibility, and human review govern scientific acceptance.
Open learning
Workshops that behave like real projects.
The OpenFOAM–ML Hackathon materials give teams a shared, executable environment for online learning, inference, and solver coupling.
Open source · hands on
Explore hackathon materials ↗OpenFOAM teaching since 2007
Reusable repositories, exercises, and live collaboration around realistic CFD + ML coupling build on long-term OpenFOAM practice and co-authorship of The OpenFOAM Technology Primer.
6completed
2ongoing
Doctoral supervision
Primary supervision with scientific ownership.
Supervision connects a precise research question, publication-quality methods, maintained software, and growing independence. Five doctoral positions have been fully funded through industry collaborations: four with Bosch Corporate Research and one with Rimac Automobili.
Additional record: ten completed master’s theses across numerical methods, data-driven CFD, optimization, and heat transfer.