KOHO reduces costs 30%, achieves zero-touch operations using EKS Auto Mode

Canada-based KOHO provides a zero-fee alternative to traditional bank accounts, expanding financial access for underserved populations. The company offers spending and savings accounts, credit building, and payment services, all of which must meet rigorous compliance requirements while being supported by a lean engineering team. 

KOHO's engineering team wanted to implement a more flexible, scalable container platform that would reduce costs, minimize operational overhead, and support future goals like multi-region resiliency and machine learning workloads. At the same time, the solution needed to maintain Payment Card Industry Data Security Standard (PCI) compliance and adhere to industry regulations. “We wanted to be able to fine-tune all our systems and clusters,” says Daniel Gienow, senior developer at KOHO. “Being able to take advantage of industry best practices using a managed service, instead of having to reinvent our own solution, was really important to us.”

KOHO also wanted to gain deeper visibility into system performance so that its engineering team could respond faster to emerging issues as the platform scaled. The company needed a solution that would let the team focus on delivering new features and improving the user experience rather than managing infrastructure complexity.

To simplify cluster management and further streamline operations, KOHO chose to migrate its clusters to Amazon Elastic Kubernetes Service (Amazon EKS) Auto Mode, which fully automates Kubernetes cluster management for compute, storage, and networking on AWS with a single click. This migration also unlocked the ability to use Amazon Elastic Compute Cloud (Amazon EC2) Spot Instances, which let users take advantage of unused Amazon EC2 capacity—resulting in a more efficient use of the underlying Kubernetes worker nodes and significant cost savings. Using these services offered the right mix of simplicity and control, making it possible for KOHO to scale its business while keeping its engineering team lean.

In December 2024, KOHO migrated its container workloads to EKS Auto Mode in just a few months. The company credits this rapid timeline in part to the operational simplicity that EKS Auto Mode provides. “Without using EKS Auto Mode, our migration effort was expected to take twice as long,” says Gienow.

Because EKS Auto Mode manages provisioning, scaling, and patching, the team modernized its infrastructure without adding operational complexity. At the same time, regulatory compliance was simplified by automating key security controls. KOHO reduced its infrastructure costs by 30 percent—despite the additional cost of EKS Auto Mode—and accelerated service startup by approximately five times. For certain services, scaling times were also reduced from 90 seconds to just 5 seconds.

Beyond faster scaling, KOHO's engineering team experienced major improvements in deployment speed. Pipelines were accelerated by 3¬–5 minutes per deployment, delivering approximately 20 percent faster deploy times. Thanks to these improvements, the engineering team can cycle through deployments and run small tests more efficiently, creating significant time savings when restarting multiple services across the platform.

Additionally, because EKS Auto Mode automates infrastructure tasks, developers don’t have to spend time fine-tuning performance, managing capacity, or monitoring instance health. Instead, they can focus on delivering new features and improving the user experience, helping the team move faster without increasing the team size. KOHO can also validate that applications run in a highly secure environment because EKS Auto Mode delivers built-in protection, automated updates, and direct integrations with other AWS security services. “Using Amazon EKS Auto Mode is improving the speed and resiliency of our system,” says Gienow. “At the same time, it’s reducing our costs significantly and removing the need to manage each individual service.”

The AWS team also provided guidance and hands-on support throughout the migration, helping KOHO stay aligned with best practices at each stage. Many of these best practices are also built into EKS Auto Mode. “It’s been incredibly helpful to tap into the deep bench of experts at AWS when we need a second opinion,” says Gienow. “Having access to people who focus on a specific area has helped us validate decisions and stay on track.”

Using EKS Auto Mode, KOHO has achieved a more scalable and cost-efficient infrastructure—and peace of mind for its engineering team. It no longer needs to manage version upgrades, installations, and security patches; instead, the team can focus on delivering new features and improving the customer experience. The company is currently working toward multi-region active configurations to support disaster recovery, a key requirement for banking compliance. 

The migration has helped KOHO progress toward its mission of helping more Canadians take control of their finances. “Using AWS, we’re able to create a modern, technology-forward financial services product that serves the needs of a larger group of users,” says Gienow. “We’re reaching communities that have been historically unserved or underserved and giving them a better financial experience.”

The new platform has also laid the foundation for long-term growth and innovation. Looking ahead, KOHO plans to use the same Kubernetes foundation to implement an internal developer platform (IDP), so engineers can create new applications quickly without managing the underlying Kubernetes infrastructure. It will also provide another platform to support KOHO’s existing machine learning and AI workloads. “AWS is an industry leader for a reason,” says Gienow. “It makes our cloud hosting experience as simple as these things get, which is extremely helpful for us.”