Are you a researcher in machine learning for speech and audio who wants to develop technology for earlier detection of lung disease? We have an exciting opportunity for a Research Associate to join the LungSight project in the School of Computer Science at the University of Sheffield.
LungSight is developing a low-cost, non-invasive audio-visual screening approach for chronic lung disease using everyday devices such as smartphones, tablets and home computers. You will lead the Sheffield machine-learning research on a large acoustic foundation model for respiratory health using real-world recordings of speech, breathing and cough. You will explore self-supervised learning, Transformer-based architectures, masked prediction and contrastive learning, with a strong focus on robustness and generalisation across speakers, demographics, environments and recording conditions.
You will curate and analyse respiratory audio datasets; develop, fine-tune and evaluate acoustic models; investigate confounding factors, background noise and demographic bias; and contribute to clinically informed acoustic modelling. A distinctive feature of the role is the close collaboration with a complementary PDRA based at the University of Southampton, whose post will focus more strongly on acoustic feature engineering, and a PDRA based at Manchester Metropolitan University, who focuses on video analysis. Together, the three posts will investigate how physiologically informed features and multi-modal data-driven representations can be combined to improve accuracy, robustness and interpretability.
LungSight is a multi-university collaboration involving the University of Sheffield, Manchester Metropolitan University, the University of Southampton, the University of Cambridge, the University of Leicester and the University of Leeds, together with NHS, clinical and third-sector partners including Asthma+Lung UK, NVIDIA, and Yorkshire Ambulance Service. You will therefore work in a highly interdisciplinary environment spanning machine learning, speech and audio processing, computer vision, respiratory physiology and clinical medicine.
You should have a PhD, or equivalent research experience, in a relevant discipline and strong practical experience in machine learning or deep learning. Experience in speech, audio or acoustic signal processing is essential. Knowledge of self-supervised learning, foundation models or Transformer-based architectures is particularly relevant. Experience in multimodal AI, healthcare or biomedical applications, or secure analysis of sensitive data would be advantageous.
The role offers excellent opportunities to publish high-quality research, develop open and reproducible research software, work closely with researchers across several UK universities, engage with end users and clinicians, and build your research career at the interface of AI and health.
We offer a highly competitive annual leave entitlement, a generous pension scheme, flexible working opportunities, and a strong commitment to staff development and wellbeing. We are committed to exploring flexible working opportunities which benefit the individual and University.
We build teams of people from different heritages and lifestyles from across the world, whose talent and contributions complement each other to greatest effect. We believe diversity in all its forms delivers greater impact through research, teaching and student experience.