Forecasting Covid Infections for the UK's NHS Using Ray and Kubernetes

Explore a 30-minute conference talk detailing the development and implementation of an Early Warning System (EWS) for forecasting COVID-19 infections in the UK's National Health Service. Learn how Faculty, a leading UK Artificial Intelligence company, created a state-of-the-art Bayesian Hierarchical Model to assess short-term risks for 219 hospital trusts. Discover the challenges faced in scaling the system, including two catastrophic outages, and how Treebeardtech resolved these MLOps issues. Gain insights into the re-architecture process that incorporated Ray Core and Kuberay, resulting in a highly stable platform for Bayesian modeling. Understand how Ray was safely integrated into a business-critical system and complemented with AWS Karpenter autoscaler. Access the comprehensive slide deck for a deeper dive into this real-world application of AI in healthcare during the COVID-19 pandemic.