DTU Skylab: Applied AI Lunch Talk

About Applied AI Lunch Talks

Applied AI Lunch Talks is an initiative designed to share insights on how AI is leveraged across various organizations to address significant challenges and create value in both business and society. Our guest speakers, who come from startups, SMEs, and large corporations, bring hands-on experience and practical knowledge about applying AI to solve complex problems within their organizations.

Talk 1

Unlocking data-driven innovation: How Lean Startup meets Agile to drive AI success at Danish Crown

AI and machine learning hold huge business potential — but adoption often creates complexity across R&D, quality assurance, and beyond. In this session, Danish Crown’s Head of Data Science shares how blending Lean Startup principles with Agile methods can unlock data-driven innovation from the bottom up. Discover how to balance structure with creativity, foster experimentation, and turn challenges into opportunities. With real-world lessons from Danish Crown’s journey, this talk offers practical insights on building an innovative culture that drives impact across industries.

Speaker

Francesc Joan Riera, Head of Data Science, Danish Crown 

Francesc Joan Riera has more than a decade worth of applied experience in computer vision and machine learning, as well as 7+ years in cloud development and project management. He has developed algorithms, ML models, and established MLOPs practices at Terma A/S, the LEGO Group, and Danish Crown A/S. 

Talk 2

AI for fetal utrasound: Every women deserves best-in-class prenatal care

More then 95% of all pregnant women take part in the ultrasound screening programme, which aims to detect and mitigate adverse pregnancy outcomes. In this talk we will discuss a research collaboration between DTU, UCPH and CAMES Rigshospitalet on AI for fetal ultrasound, as well as the journey from research, via spinout, to clinic.

On the research side we will discuss how different types of models give different types of support to the clinicians. We will discuss how to create explainable AI tools to aid clinicians' learning and clinical decisions; we discuss how different types of models support different types of tasks; and we discuss the biases found in different types of models. On the innovation side we discuss the journey and callenges encountered in moving from research to a spinout that creates real-life AI products. 

Speaker

Aasa Feragen, Professor at DTU Compute & P1 Co-Lead

Professor Aasa Feragen is at DTU Compute and P1 co-lead of the Causality & Explainability collaboratory. Her research integrates medical imaging, fairness, and explainability. She co-founded Prenaital, a spin-out leveraging AI to detect high-risk pregnancies earlier and improve prenatal diagnostics.