01 / Mechanistic learning Python
MxlPy
Connect mechanistic modeling with machine learning to build explainable, data-informed models.
Research software engineer / RWTH Aachen
Hello there π
Scientific ideas.
Working software.
Iβm Marvin. I build scientific software for universal differential equations, connecting mechanistic models with machine learning.
My research background is in biological modeling and photosynthesis. I like making complex models easier to work with.
Universal differential equations
dudt=f(u, p, t)+UΞΈ(u, t)
Known dynamicsLearned component
01 / Selected software
All software β02 / Dynamic modeling Python
modelbase
Build and analyze dynamic mathematical models of biological systems, from reactions to metabolic networks.
03 / Reusable building blocks Python
MxlBricks
Build mechanistic learning models quickly from re-usable reaction bricks on top of MxlPy.
02 / Research behind the tools
All papers β03 / A little background
Biology brought me here.
Building keeps me curious.
I studied Biology and Quantitative Biology at Heinrich Heine University DΓΌsseldorf, where I completed my PhD on computational analysis and optimisation of photosynthetic carbon fixation.
Today, I work as a postdoc and research software engineer in the Computational Life Science lab at RWTH Aachen.
Talks & teaching β RWTH Aachen University
Computational Life Science