Research Associate in Speech and Acoustic Sensing for Respiratory Health

This position at the University of Sheffield offers a Research Associate role within the LungSight project, focusing on advancing machine learning techniques for respiratory health. The successful candidate will lead Sheffield’s machine-learning research on a large acoustic foundation model for respiratory health, leveraging real-world recordings of speech, breathing, and cough. The role involves developing self-supervised learning models, Transformer-based architectures, and robust multimodal AI systems to improve early detection of lung diseases. Collaborative work with researchers at multiple UK universities and clinical partners will ensure the integration of physiologically informed features and multi-modal data-driven representations for enhanced accuracy and interpretability.

• Develop and fine-tune large acoustic foundation models for respiratory health using real-world speech, breathing, and cough recordings.
• Explore self-supervised learning, Transformer-based architectures, masked prediction, and contrastive learning techniques.
• Curate and analyze respiratory audio datasets, focusing on robustness and generalization across diverse demographics and environments.
• Investigate confounding factors such as background noise and demographic bias in acoustic models.
• Collaborate with researchers at the University of Southampton and Manchester Metropolitan University to integrate multimodal data-driven representations.
• Contribute to clinically informed acoustic modeling and work closely with NHS and clinical partners.
• Develop open and reproducible research software and publish high-quality research outcomes.