I’m a PhD student in Electrical and Electronic Engineering at Imperial, and part of NeuroWare, the national Innovation and Knowledge Centre in Neuromorphic Computation. I work across machine learning, computational neuroscience and digital hardware design: training delay-based spiking models, studying the trade-offs between temporal representation, weight precision and memory, and mapping these models onto FPGAs with hardware-aware measures of energy, memory and latency. More broadly, I'm interested in intelligence and its physical implementation: what can be learned and computed under tight power and memory budgets, as on edge devices and event-based sensors.

Carl is supervised by George Constantinides and co-supervised by Dan Goodman and Christos-Savvas Bouganis.