Research Associate in Learning Algorithms and Simulation Spintronics Devices

Research Associate in Learning Algorithms and Simulation for Spintronics Devices

We are seeking a motivated and collaborative individual to join our team as a Research Associate in Learning Algorithms and Simulation for Spintronics Devices. This role offers an exciting opportunity to contribute to the SkyANN, an interdisciplinary project on Skyrmionic Artificial Neural Networks, focused on exploring early-stage concepts for brain-inspired hardware using emerging spintronic technologies, within a dynamic and inclusive environment.

• Develop neural network models informed by experimental and simulated skyrmionic device behaviour, capturing key physical characteristics relevant to learning and inference.
• Translate device-level properties, such as noise, variability, stochasticity, dynamic response, locality, nonlinearity, and limited precision, into computationally efficient neural network and compact model abstractions.
• Design and implement hardware‑aware training and inference workflows in modern machine learning frameworks, such as PyTorch, Tensor Flow, JAX, or similar.
• Explore physics‑inspired and brain‑inspired learning methods suitable for constrained neuromorphic hardware, including local learning rules, surrogate models, adaptive learning, error tolerance, quantisation, robustness methods, and unsupervised or semi‑supervised learning where appropriate.

We welcome candidates who bring diverse perspectives, experiences, and approaches to their work.

We encourage applications from individuals with a wide range of backgrounds and experiences. You should demonstrate:

• PhD awarded or near submission in a relevant discipline, such as computer science, machine learning, physics, electrical and electronic engineering, applied mathematics, computational neuroscience, or a related area.
• Strong experience in neural networks and modern machine learning methods, with evidence beyond basic coursework or routine applied use.
• Proven ability to implement, modify, train, and evaluate neural network models in an ML framework, such as PyTorch, Tensor Flow, JAX, or similar.
• Strong computational and programming skills, with evidence of delivering clean, testable, research‑quality code and reproducible workflows.
• Ability to abstract physical, dynamical, or device‑level behaviour into computationally efficient models suitable for neural network training, inference, or benchmarking.

• Hardware‑aware training or inference for analogue, in‑memory, neuromorphic, memristive, spintronic, or other emerging computing hardware.
• Learning algorithms designed for limited resources, scarce labelled data, noisy updates, device variability, low precision, or non‑ideal hardware behaviour.
• Unsupervised, self‑supervised, semi‑supervised, local, biologically‑inspired, or adaptive learning methods.
• Custom neural network layers, surrogate models, differentiable device models, or physics‑informed neural network components.
• Numerical simulation of dynamical systems, including ODEs, PDEs, nonlinear dynamics, or dynamic synapse models.

We value transferable skills and real‑world experience as much as formal qualifications.

• Generous employer contribution pension.
• 29 days annual leave plus bank holidays, along with Christmas closure.
• Ride to work and EV car scheme available.