How Stone runs payments and banking on C#, Go and three clouds
Stone sells card machines, a digital business account and credit to Brazilian merchants. Its engineering splits by business line, with C# and .NET in credit, Go in banking, Kotlin Multiplatform on mobile, and infrastructure spread across AWS, Google Cloud and Azure.
Contact center that puts every customer interaction in one view, including WhatsApp.
Why: Stone saw it as the most customizable platform on the market, with global reach.
How they built it
Two backend languages, split by business
Stone does not publish an engineering blog, so its current openings are the clearest public picture of the stack. The credit team builds in C# with Docker, Azure DevOps pipelines and Datadog or Elasticsearch for observability, and asks for event sourcing, CQRS and hexagonal architecture.
The banking teams behind boletos, account management and investments hire for Go, working with relational databases, messaging and streaming, and AI-assisted tools like Claude Code or Codex.
The Foundation team supports internal customers on AWS, Google Cloud and Azure at once. Its work covers VMs, networking, DNS and CDN, delivery pipelines in GitHub Actions and Azure DevOps, cloud cost management, and moving workloads between clouds and from on-premises to the cloud.
Infrastructure is written as code with Terraform and Ansible, and workloads run in Docker on Kubernetes and ECS.
On mobile, Stone shares business rules between Android and iOS with Kotlin Multiplatform, and lists it as a must-have for senior mobile engineers. Screens use Jetpack Compose, Compose Multiplatform or SwiftUI.
The web side of the same products is built in React and TypeScript.
Stone moved its contact center to Twilio Flex to see every customer interaction in one place, then added a WhatsApp integration and AI on top. It picked Twilio as the most customizable platform with global reach, and reported higher customer satisfaction.
A GenAI team in Florianópolis now builds generative AI and machine learning features for these conversational products on WhatsApp and Twilio, mostly in Python and TypeScript.
