SquareOps
Get an architecture review

How SquareOps Took DexKor From Zero Infrastructure to Production on AWS EKS in One Week

SquareOps Team October 28, 2026 Case Studies

The Challenge

DexKor had to ship a multi-tenant enterprise platform on Kubernetes while the whole team was focused on the product, not on infrastructure.

  • No in-house DevOps. The founding team are product and engineering people, and hiring a DevOps engineer was not viable at their stage.
  • Dedicated tenant environments. Each enterprise customer needed an isolated deployment with its own resources, data separation and configuration.
  • Speed to market. Competing for enterprise deals meant weeks spent on infrastructure were weeks of lost revenue.
  • Continuous delivery. Enterprise customers expect regular updates and quick fixes, including at weekends, without downtime.

What SquareOps Did

SquareOps set DexKor up on its Atmosly internal developer platform, so developers could own the full deployment lifecycle without learning the details of Kubernetes, Helm, networking and cloud IAM.

  • A production AWS EKS cluster in days. Atmosly provisioned the EKS cluster with its networking, security groups and IAM roles in place. DexKor went from zero infrastructure to a live production environment in one week.
  • Environment cloning for each tenant. Onboarding a new enterprise customer used to be a multi-day DevOps task. Cloning a configured environment turned it into a one-click operation, so the team could create an isolated tenant environment, with its services and configuration, in minutes and without writing Kubernetes manifests.
  • CI/CD from day one. Atmosly pipelines build, test and deploy automatically. Developers push code and the pipeline handles container builds, the image registry and rolling deployments, with rollback options when needed.
  • Visibility across tenants. The team could see what was running in each tenant environment and which pipeline run had put it there, without assembling a monitoring stack of their own.

Results

  • Zero to production in one week, with the EKS cluster, CI/CD pipelines and the first tenant environment live.
  • New tenant environments in minutes, which removed the bottleneck in onboarding enterprise customers.
  • Multiple deployments a day, seven days a week, with no DevOps gatekeeping.
  • No DevOps hire needed. The same developers who build the product run the infrastructure through the platform.

Conclusion

DexKor shows that an early-stage SaaS team does not need a DevOps team to run production Kubernetes. With SquareOps and Atmosly handling cluster setup, environment management and CI/CD, a team of under ten ships updates daily to enterprise customers on a platform that grows with each deal.

Frequently Asked Questions

How long did it take DexKor to reach production?

One week. SquareOps and Atmosly provisioned the AWS EKS cluster, set up the CI/CD pipelines and brought up the first tenant environment within that time.

How does DexKor give each enterprise customer an isolated environment?

Each customer gets a dedicated environment created by cloning a configured one. It carries its own services and configuration, and takes minutes rather than the multiple days it used to take.

Does DexKor need a DevOps engineer?

No. The team of under ten manages deployments and environments through Atmosly, so the developers building the product also run the infrastructure.

How often does DexKor deploy?

Several times a day, including weekends, through automated build, test and deploy pipelines.