Research and Innovation Projects

University of Sheffield, UK

• Develop scalable AI systems to enable early identification of dementia, mild cognitive impairment, depression, anxiety, motor neuron disease, and post-stroke cognitive decline (CognoSpeak, CognoMND and CognoStroke).
• Establish clinically interpretable speech and behavioural biomarkers supporting objective cognitive assessment and clinical decision-making.
• Curate and quality-control longitudinal multimodal datasets from 5,000+ NHS participants in collaboration with clinicians to enable real-world clinical AI research.
• Investigate relationships between attention, memory, and executive function measures and speech-derived behavioural markers of cognitive decline.

Stellenbosch University, South Africa

• Develop non-invasive, low-cost AI screening approaches for tuberculosis and respiratory diseases using cough audio collected in real clinical environments across South Africa and Uganda.
• Build large-scale cough datasets supporting robust machine learning model development for global health diagnostics.
• Contribute to public release of rare clinical audio datasets to accelerate research in AI-based respiratory disease detection at the end of the CAGE-TB project.

UNLV, USA

• Modification detection: Study the relationship of base modifications to disease pathology at the genomic/transcriptomic scale using Oxford Nanopore sequencing and deep learning.
• Long-read data analysis: Develop computational tools via machine learning especially deep learning for genomic and transcriptomic data analysis on long-read sequencing data.

Stellenbosch University, South Africa

• Develop machine learning systems analysing cough, breath, and speech signals for rapid remote COVID-19 detection.
• Model disease severity and hospitalisation risk using large-scale international clinical audio datasets.

Stellenbosch University, South Africa

• Automatic detection of coughs among other environmental noises in a multi-bed ward at TASK (a 24h TB research centre).
• Enable automated cough detection and bed occupancy monitoring in hospital wards to support longitudinal healthcare analytics using a non-invasive smartphone-inbuilt accelerometer.

Stellenbosch University, South Africa

• Develop robust machine learning models distinguishing TB coughs from other respiratory conditions under noisy real-world recording conditions.

University of Edinburgh & Stirling, UK

• Design and develop a multi-channel adaptive MEMS/CMOS microphone system inspired by human auditory processing.