Jan Vogelsang

I am a PhD student specializing in high-performance computing for computational neuroscience, bridging the gap between brain-scale simulation and hardware optimization. I am a core developer of the NEST spiking neural network simulator, where my recent work on its core algorithms has yielded a 50% performance improvement.

My current research interest is software-hardware co-design. I am exploring methods to automatically discover optimal algorithms tailored to specific hardware architectures. The goal is to move beyond the current manual, trial-and-error process of ML development by automating the optimization of algorithms, architectures, and hardware targets. This could involve systems that generate abstract algorithm representations or leverage AI-assisted compilers to optimally transform these algorithms into efficient, hardware-specific code. I aim to apply this approach not only to accelerate neural simulations but also to broader ML applications, ultimately creating a new paradigm for evaluating and co-developing novel neuromorphic hardware.