AWS

AWS dominates the cloud infrastructure market for good reason: its breadth of services lets platform teams assemble opinionated internal platforms without building everything from scratch. EKS for managed Kubernetes, CodePipeline and CodeBuild for CI/CD, CloudFormation and CDK for infrastructure-as-code, and IAM for fine-grained access control form the backbone of most enterprise platform engineering stacks. The ecosystem is deep enough that nearly any operational pattern has a managed-service answer.

Platform engineers working with AWS spend significant time on IAM policy design, VPC networking, and service quotas—the unglamorous connective tissue that determines whether a self-service platform actually works at scale. Getting cross-account access right with Organizations and Control Tower, wiring up PrivateLink endpoints, and tuning autoscaling policies across EKS node groups are where real operational expertise lives.

The tradeoff is complexity. AWS offers multiple ways to accomplish any goal, and choosing between them has long-term consequences for cost, maintainability, and team cognitive load. A well-built AWS platform abstracts that complexity behind golden paths so application teams get the reliability of battle-tested infrastructure without needing to understand every service interaction underneath.