Research ThemeNovel Computing and Beyond-CMOS hardware AimTo develop reconfigurable magnetic reservoirs that exploit the nonlinear dynamics of advanced magnetic materials for ultra-low-power, real-time in-memory computing and classification of complex sensor signals, targeting practical edge applications including biosensing and sonar technologies. ObjectivesDevelop and characterise magnetic reservoirs based on advanced magnetic materials and topological spin textures, prioritising material platforms that minimise reliance on critical raw materials and heavy metals to support scalable, sustainable and supply-chain-resilient technologies. Design reconfigurable in-memory computing architectures in which magnetic states and device parameters can be tuned to optimise temporal information processing, establishing quantitative relationships between material properties, spin dynamics and computational performance. Demonstrate real-time processing of complex temporal sensor signals across complementary applications: (i) electrocardiogram (ECG) heartbeat classification as an accessible biosensing benchmark; and (ii) sonar digital signal processing, developed with KAON Ltd, targeting relevant tasks such as signal/event detection, classification and temporal discrimination under challenging or noisy conditions. Optimise magnetic reservoirs for practical edge applications using multiscale modelling, and machine-learning-assisted design, targeting low energy consumption compact hardware and high classification performance relative to conventional computing approaches.DescriptionArtificial intelligence is increasingly moving from large computing centres towards sensors and edge devices, where conventional hardware faces severe constraints in energy consumption, size and data movement. Reservoir computing offers a radically different approach: the nonlinear dynamics of a physical system can transform and process time-dependent information directly, potentially reducing reliance on energy-intensive digital computation. This project will develop reconfigurable magnetic reservoirs using advanced magnetic materials and topological spin textures, such as skyrmions, whose nonlinear and history-dependent dynamics provide a natural platform for in-memory computing. Through multiscale simulations combined with machine learning, the student will establish how material properties, magnetic interactions and device architecture determine computational performance, while prioritising materials with reduced dependence on critical raw materials and heavy metals. The technology will be evaluated through two complementary demonstrators. Heartbeat classification from electrocardiogram (ECG) signals will provide an accessible biosensing benchmark. In collaboration with KAON Ltd, a second demonstrator will investigate magnetic-reservoir approaches to sonar digital signal processing, including detection, classification and temporal discrimination of signals under challenging conditions. The project will establish a pathway towards compact, adaptive and ultra-low-power intelligent sensing hardware, with applications spanning wearable monitoring, sonar processing and autonomous defence systems where energy, latency and computational resources are constrained. Closing date: 31 Jan, 2027 Apply now Principal Supervisor Dr Elton Santos Assistant Supervisor Prof Rebecca Cheung Eligibility Minimum entry qualification - an Honours degree at 2:1 or above (or International equivalent) in a relevant science or engineering discipline, possibly supported by an MSc Degree. Further information on English language requirements for EU/Overseas applicants. Funding This position is fully funded. Informal Enquiries spads@ed.ac.uk
