Post-Doctoral Research Associates
The School of Informatics at the University of Edinburgh invites applications for several Post-Doctoral Research Associates (PDRAs) to conduct research on novel architectures for Large Language Models. Successful candidates will work under the supervision of Professors Frank Keller, Mirella Lapata, Amos Storkey, and Ivan Titov—leaders in NLP, machine learning, and cognitive modeling—and collaborate closely with partner institutions.
Five positions are part of the Science of Fundamental AI Research (SOFAIR) Lab, a national AI research partnership between UCL and the Universities of Cambridge, Edinburgh, and Oxford, funded as part of a £60 million investment from UK Research and Innovation (UKRI). SOFAIR brings together experts across computer science, mathematics, statistics, and neuroscience to explore new AI architectures designed to run on widely available hardware, making cutting-edge AI broadly accessible.
The PDRAs will be responsible for developing fundamental AI modeling and learning methods beyond standard transformers and next-token prediction. Research areas include modular, compositional, and compute-adaptive models, neurobiology-inspired architectures, sparse/structured reasoning, and new training paradigms such as reinforcement-learning hybrids and gradient-free methods. A primary goal is to improve reasoning, efficiency, and interpretability while enabling training and inference across heterogeneous, distributed, small-memory hardware.
Key duties and responsibilities include:
Designing and implementing novel model architectures and scalable training algorithms.
Conducting controlled experiments to evaluate reasoning, efficiency, interpretability, and generalization.
Training and benchmarking large models using multi-GPU and distributed-computing infrastructure, including access to Isambard-AI, the UK’s national AI supercomputer with over 5,000 Nvidia GH200 GPUs.
Developing robust research code, analyzing model behavior, and disseminating findings through high-impact publications and presentations.
These roles offer funding for international conference travel and a dedicated compute allocation on Isambard-AI. The positions are open to UK and international applicants, with visa sponsorship available. The post is advertised as full-time (35 hours per week, fixed-term for 14 months), but part-time or flexible working patterns and hybrid working requests (a mix of remote and regular on-campus work) will be considered.
A PhD (or nearing completion) in Computer Science, Informatics, Artificial Intelligence, Machine Learning, Computational Linguistics, or a closely related quantitative field.
Open to both UK and international applicants (visa sponsorship is available).
Proven research expertise in machine learning, natural language processing, deep learning, or computational cognitive modeling.
Strong programming skills in Python and deep learning frameworks (e.g., PyTorch, JAX).
Experience with designing, training, and benchmarking large-scale machine learning or language models on multi-GPU/distributed infrastructure.
Strong background in qualitative and quantitative analysis of model behavior, reasoning, efficiency, and interpretability.
Demonstrated ability to write robust research code and publish/present research findings in top-tier international venues.
£41,064 to £48,822 per annum (Grade UE07)
September 7, 2026 (Job Ref: 14684)