Celina Caroto

Celina Caroto is a PhD researcher and student in the Department of Computer Science & Information Systems at the University of Limerick, specialising in probabilistic and explainable deep learning for medical diagnostics and clinical decision support. She holds a postgraduate degree in Computer Science (with a dissertation on AI-driven predictive maintenance for safety equipment) and a Bachelor’s degree in Computer Science (with a dissertation on AI-based endometriosis diagnostics), both from the Richfield Graduate Institute of Technology. Her research focuses on multimodal AI pipelines that combine medical imaging and structured clinical data, with a particular emphasis on computer vision, Bayesian deep learning, and uncertainty quantification to support non-invasive diagnosis of endometriosis and adenomyosis. A core strand of her work is the design of clinician-centred explainability and trust mechanisms, ensuring that model outputs, explanations, and uncertainty estimates are transparent, interrogable, and aligned with safety, ethical, and regulatory requirements in high-stakes care.

In parallel with her academic work, Celina has extensive industry experience across AI-integrated systems, automation engineering, and software quality in regulated, safety-critical environments. She has designed and refined end-to-end automation frameworks, validated machine learning workflows, and optimised CI/CD pipelines for complex, large-scale enterprise platforms. Her industry background includes contributions to organisations such as Google, Johnson & Johnson, the NHS, Baxter Healthcare, and leading SaaS and medical technology companies, where she has delivered backend infrastructure, automation pipelines, and data-driven QA strategies for clinical, diagnostic, and NLP-based systems. She specialises in AI model validation, microservices and API testing, and performance benchmarking, leveraging her ongoing research in multimodal diagnostic AI and uncertainty-aware evaluation to bridge cutting-edge methods with robust, trustworthy deployment in real-world settings.

Explainable AI (XAI), probabilistic and Bayesian deep learning, uncertainty-aware medical imaging, Computer Vision for medical diagnostics, multimodal fusion of imaging and structured clinical data, AI for women’s health (endometriosis and adenomyosis), trust and adoption of AI in clinical workflows, federated and privacy-preserving learning in healthcare.