From Raw Data to Real Production: Unlock High-Stakes AI w/ Quality Playground— Erik Duhaime, Centaur скачать в хорошем качестве

From Raw Data to Real Production: Unlock High-Stakes AI w/ Quality Playground— Erik Duhaime, Centaur

Recorded live on February 12, 2026, this episode of Bytes of Innovation features an exclusive fireside chat exploring one of the most critical — and often overlooked — drivers of AI success in healthcare: quality. In this session, Erik Duhaime, PhD, CEO and Co-Founder of Centaur, shares a practitioner’s perspective on building high-stakes AI systems for medical imaging and healthcare. Titled “From Raw Data to Real Production: Unlocking High-Stakes AI with the Quality Playground,” the conversation examines why model performance alone is not enough — and how data quality, labeling accuracy, and human expertise ultimately determine whether AI systems succeed in real clinical environments. Drawing on his experience developing AI infrastructure used by leading healthcare organizations, Erik discusses how AI teams operationalize quality across the full lifecycle — from training and validation to deployment and continuous improvement. 🔹 Hosted by Aline Lutz, MD PhD – VP of RWE and Clinical Insights, Segmed Martin Willemink, MD PhD – Chief Scientific Officer & Co-founder, Segmed 🎥 What You’ll Learn Why data and labeling quality remain the biggest bottlenecks in healthcare AI How leading AI teams define and measure quality beyond traditional performance metrics The role of human-in-the-loop processes in improving model reliability and trust Common pitfalls that degrade AI reliability in medical imaging workflows How strong quality systems enable safer, more scalable AI adoption Why quality must be treated as a strategic priority — not a downstream task 💡 Why This Matters As AI adoption accelerates across radiology and medical imaging, expectations for safety, performance, and trust continue to rise. Yet many AI initiatives struggle due to inconsistent datasets, unclear evaluation standards, and insufficient quality controls. This discussion reframes quality as the foundation of production-ready AI. Rather than focusing solely on algorithm performance, it highlights how sustainable, real-world impact depends on building rigorous quality processes into AI systems from the ground up. 👉 Explore Bytes of Innovation Visit http://bytesofinnovation.ai and stay updated on upcoming episodes. 🔔 Subscribe to Segmed on YouTube for more conversations with leaders shaping the future of medical imaging and responsible AI.