Case Study: Learning Machine cuts code from 25,000 to 500, accelerates deployments and halves AWS costs with Pulumi

Learning Machine Expediting Customer Deployments

Learning Machine, a provider of blockchain-secured digital records, faced fragile, hard-to-maintain infrastructure-as-code: 25,000 lines of ad-hoc templates and scripts that only one DevOps engineer fully understood, slowing deployments and blocking CI/CD. To meet a customer-driven need for single-tenant application suites and improve developer agility, Learning Machine adopted Pulumi’s cloud-native infrastructure-as-code on AWS.

Pulumi helped Learning Machine refactor and replace the sprawling scripts with TypeScript-based Pulumi code (one month of planning and advisory), enabling containerized deployments on Amazon ECS and AWS Fargate. The result: the codebase shrank from 25,000 lines to 500 lines of JavaScript, deployment time fell from one month to under a day, AWS costs were cut in half, and Learning Machine gained CI/CD, faster provisioning, and more reliable, maintainable operations — all delivered with Pulumi.

Learning Machine

Dan Hughes

Chief Operating Officer