PhD Studentship in Neuromorphic Engineering

The PhD project will focus on the design of analogue and mixed-mode processor circuits for implementing computing systems, in particular accelerators for machine learning (deep learning, neural networks). The PhD student will be associated with the FORTE project (www.forte.ac.uk), which is a collaboration between The University of Manchester (Schools of EEE and Computer Science), Imperial College London, and University of Southampton. The project aims to develop configurable circuits using memristors – novel nanoelectronic devices which promise to revolutionise the design of future computing systems.  

The student will investigate the methods of mapping computations onto suitable circuit topologies, implementation, and system-level aspects of the design. Of particular interest is the question of exploiting device characteristics, and the variability and stochastic nature of the device operation. The project will involve prototyping custom VLSI integrated circuits, which will be designed using advanced EDA tools for circuit simulation and layout, fabricated using state-of-the-art commercial CMOS fabs, and integrated with novel memristor devices. Complete devices will be tested using our comprehensive set of IC test equipment.

The candidates should have a UK first class honours degree (or equivalent) in electronics, electrical engineering, computer engineering or a similar subject. They should have a strong interest in both hardware circuit design and novel computing architecture, with a motivation to pursue a doctorate in an intellectually challenging field requiring both the originality of thinking as well as developing solid hardware design skills.

Interested applicants should send an up-to-date CV to Professor Piotr Dudek ([Email Address Removed]) with a cover letter explaining their interest and suitability for this research. The project is available immediately, and the start date will be arranged to coordinate with the admissions cycles and the candidate’s availability.