Verivox, a leading German online comparison platform, has successfully modernized their AI/ML workflow on AWS to stay at the edge of the industry. We will explore how Verivox transitioned from an on-premise Machine Learning workflow using R and proprietary tooling, to migrating to a cloud setup with Python and Amazon SageMaker while implementing MLOps best practices (CI/CD, IaC via CDK). This transition enabled data scientists to own the full stack of the ML Workflow, which drastically reduced their Model Development and Deployment time from Weeks to Hours.