I build computational tools that help people make better decisions in complex, uncertain worlds—from urban climate planning to molecular design.

My work focuses on Evolutionary Computation, Quality Diversity (QD), and Surrogate-Assisted Optimization. I treat AI as a co-designer: a partner that helps humans jointly explore, understand, and learn from extremely large data and optimization domains.

• Languages: Python (Primary), C++, MATLAB, Java, Shell.
• AI & ML: Surrogate-assisted Multi-Solution Optimization (SPHEN), Gaussian Processes, Neuroevolution (NEAT/HyperNEAT), Quality Diversity (QD), Deep Generative Models (GANs, U-Nets), Agentic AI.
• Cheminformatics & Bio-AI: Variant Effect Prediction, Molecular Conformer Prediction, Virtual Screening, Force Field Parameterization.
• Scientific Computing: Computational Fluid Dynamics (CFD), Shape Optimization, Lattice-Boltzmann Methods.
• Tools: PyTorch, Scikit-learn, ROS, Git, LaTeX, Docker.

• Variant Effect Prediction: Developing models to predict the functional impact of genetic variations.
• Molecular Design: Applying generative modeling and optimization for Prediction of Molecular Conformers and high-throughput Virtual Screening.
• Force Field Modeling: Using surrogate-assisted global optimization to refine Lennard-Jones parameters and molecular system accuracy.
• CytoTransport (DFG): In-silico modeling of cellular transport mechanisms at the intersection of biomedicine and structural biology.

• OpenSKIZZE: (DBU-funded) An open-source AI assistant for climate-adaptive urban design. It uses surrogate-assisted QD to generate thousands of diverse building layouts evaluated for cold airflow impact without live CFD simulations.
• Digital Twins: Representative for H-BRS at the GeoIT Round Table NRW and contributor to Digital Twins 4 Multiphysics (DT4MP) labs.

• sphenpy: A Python library for Surrogate-assisted Phenotypic Niching. It bridges the gap between high-quality optimization and expensive scientific simulations.
• v-elites: MATLAB implementation of Voronoi-Elites and prototype discovery methods for divergent thinking.
• haneat-gecco2017: Neuroevolution research into evolving parsimonious networks with mixed activation functions.

• Strategic AI: Applying AI to reverse the engineering cycle—showing designers early on which types of solutions meet complex climate or fluid dynamics requirements.
• Awards: GECCO Best Paper Award (Honorable Mention 2023, 2017), ACM SIGEVO Best Dissertation Award (Honorable Mention 2022), RoboCup Symposium Best Paper (2016).
• Community: Active in the Neue Stadtgärtnerei (Bonn-Dransdorf), a grassroots initiative for climate-resilient urban spaces.