There, he was fellow and cluster speaker of the transdisciplinary NRW research college Light, Efficient, Mobile and started working in the intersection of uncertainty quantification and micromechanics, in particular computational homogenization, using deep learning approaches. Furthermore, he worked on physics informed neural networks to directly solve the underlying system of PDEs.

In 2021, he switched to TU Braunschweig and proceeded with his PhD under Prof. Dr. Henning Wessels at the Institute for computational modeling in civil engineering (iRMB). At TUB he worked on the generation of microstructures using generative adversarial neural networks. He completed his PhD in 2022 and currently works on brain-inspired so-called spiking neural networks which are heavily studied in computational neuroscience. They can utilize brain-inspired neuromorphic chips, which are magnitudes more energy efficient than current hardware. Together with Prof. Eshraghian from the University of California, Santa Cruz, a distinguished developer of neuromorphic chips, he published a recent paper on nonlinear, history-dependent regression using spiking neural networks. Currently, he works at ETH Zürich.