DEIMinari: From Pixels to Pandemics: An AI Journey Across Academia, Industry and Back to Global Health

Títol: From Pixels to Pandemics: An AI Journey Across Academia, Industry and Back to Global Health

Conferenciant: Md Mostafa Kamal Sarker

Institució: Big Data Institute, University of Oxford

Professor/a organitzador/a: Domènec Savi PUIG VALLS

Hora: 12:00h

Resum:

In this talk, Dr Mostafa Sarker traces his 14-year journey from medical imaging research through industry and back to academia, where he now leads AI for global health at the University of Oxford. He explores how AI methods can be transferred between fields and why their practical impact depends on people, data and partnerships. His academic work began with doctoral research in Spain on recognising food places in egocentric photo-streams from wearable cameras and extended to skin lesion segmentation for skin cancer diagnosis. He subsequently worked on computational pathology within PathLAKE at Queen's University Belfast, and on underwater object detection and AI for thermographic surveys at Robert Gordon University. At Oxford's Department of Engineering Science, he developed clinical AI for detecting congenital heart defects in fetal echocardiography within the InnoHK-funded COCHE centre. As Head of AI at Technovative Solutions and Consortium AI Task Force Lead, Dr Sarker wrote the AI strategy and work packages for three funded Horizon Europe projects: CLIMATEADAPT4EOSC on climate adaptation, EcoPlast on circular automotive plastics, and GENMAT on generative foundation models for materials discovery. These projects expanded his work on domain-specific foundation models and agentic AI beyond healthcare. He now leads the AI in Global Health group at MORU Epidemiology and serves as Oxford's AI Task Force Lead for PANDAI, building the foundation model for the European Pandemic AI Observatory, with the WHO Global Hub for Pandemic and Epidemic Intelligence as a key partner and end user. The talk concludes with practical lessons on moving between domains, developing diverse AI solution systems, and building trustworthy, deployable AI for low-resource and global health settings.