at CPS-ME leading research on DIRC: Dynamics-Informed Reservoir Computing for DFG SPP 2353:
- PyReCo - Designed and released the open-source PyReCo library.
- Node Pruning - Algorithm that reduces network size while improving performance. Publications in Chaos 2025. Featured in AIP Scilight.
- PDNE - Evolutionarily-guided efficient & sparse network framework outperforming large ML architectures (PRE 2025).
- Digital Twins - for predicting nonlinear oscillator bifurcations and chaos from minimal sensor data (Nonlinear Dyn. 2024).
- Physical ML Device - running on water ripples with a Raspberry Pi readout layer for energy-efficient, hardware-native computation.
- Supervised 10+ theses in ML. Invited keynote at GACM 2025. Co-organizing RCC26, Berlin.