What you'll learn

Cloud architects are increasingly handed AI requests before anyone has defined what the workload actually needs, which leads to designs that are either over-engineered and expensive or under-built and ungoverned. In this course, AI and ML Workload Considerations for the Cloud Architect, you’ll gain the ability to determine the minimal engineering any AI or ML workload requires by working it as a solution-architecture problem. First, you’ll explore how to translate a workload description into the compute, data, identity, and lifecycle characteristics it implies, and how to recognize workloads that belong outside foundation-model architecture. Next, you’ll discover how to decompose a workload into the competing forces of performance, cost, rigor, and operational burden, and how to establish the rigor criteria an architecture must satisfy before any platform or service is considered. Finally, you’ll learn how to derive the minimal architecture a workload requires and justify why additional components would be over-engineering. When you’re finished with this course, you’ll have the skills and knowledge of AI and ML workload architecture needed to determine the minimal engineering any AI or ML workload requires.