Data-Driven Surrogate Modeling
High-fidelity simulations of crash, structural, and musculoskeletal systems are too slow for design loops, control, or real-time use. My PhD work (University of Stuttgart, SimTech) built data-driven surrogates that learn a low-dimensional latent representation of the full-order solution and model its dynamics there — without access to the simulation code.
Highlights include non-intrusive nonlinear model reduction for structural dynamics, multi-hierarchical surrogates combining graph neural networks with mesh simplification, real-time human response prediction, and surrogate-accelerated stability analysis. The work is summarized in my dissertation.
Jonas Kneifl
Postdoctoral Researcher
My research combines model order reduction, surrogate modeling, and machine learning — and asks what physics generative models learn.
