{"slug":"deutsche-bahn-runs-a-shared-ai-platform-on-amazon-sagemaker-studio-for-over-100-developers","url":"https://findausecase.com/use-cases/deutsche-bahn-runs-a-shared-ai-platform-on-amazon-sagemaker-studio-for-over-100-developers","title":"Deutsche Bahn runs a shared AI platform on Amazon SageMaker Studio for over 100 developers","description":"Deutsche Bahn's AI platform team (DB Systel) operates Amazon SageMaker Studio as a group-wide platform where multiple data analytics teams develop, train and run ML workloads, for use cases ranging from railway maintenance and forecasting to future generative AI applications. Each team requests a SageMaker domain through a self-service portal, and the VPC has no outbound internet connectivity. A product owner says that within 8 weeks over 120 developers were onboarded and 25 SageMaker domains provisioned; the post's conclusion cites over 100 developers and 20 domains.","company":"Deutsche Bahn","industry":"Transportation","country":"Germany","aiCapabilities":["Machine Learning"],"technology":["Amazon SageMaker"],"deployment":"Public Cloud","problemStatement":"Before the shared platform, each Deutsche Bahn data analytics team managed and operated its own JupyterLab notebooks, which was not efficient or cost-effective; the organization needed a scalable platform that meets compliance and security standards.","solutionApproach":"A dedicated AI platform team manages Amazon SageMaker Studio from a central platform account. Each data analytics team requests one or more SageMaker domains through an internal self-service portal; an orchestration workflow provisions the domain, an Azure AD group, an IAM role, an S3 bucket and a customer managed key through a CI/CD pipeline deploying an AWS CDK app. The VPC has no outbound internet connectivity, and lifecycle configuration shuts down idle instances to reduce unnecessary costs.","businessValue":"According to the DB Systel product owner, SageMaker Studio provided a common platform that is scalable and security compliant, and within 8 weeks over 120 developers were onboarded and 25 SageMaker domains provisioned. The post's conclusion cites over 100 developers and 20 domains in a single AWS account, and names high scalability through automation and self-service plus a consumption-based pricing model as major benefits.","evidence":{"band":"high"},"sourceUrl":"https://aws.amazon.com/blogs/machine-learning/supercharge-your-ai-team-with-amazon-sagemaker-studio-a-comprehensive-view-of-deutsche-bahns-ai-platform-transformation/","dates":{"publishedAt":"2026-10-11T05:46:58.279Z","publishedAtSource":"pipeline","updatedAt":"2026-10-11T05:46:58.279Z"},"license":"Open for reading and citing with a link to https://findausecase.com/use-cases/deutsche-bahn-runs-a-shared-ai-platform-on-amazon-sagemaker-studio-for-over-100-developers. Bulk republication requires permission."}