I build the cloud platforms other engineers deploy onto.

Geotab, Senior Platform Engineer, 2019 - present

Leaving the data center for GKE

Problem

A routing and optimization domain ran on legacy hosting split between on-premises data centers and AWS, as larger applications that were hard to scale and release independently.

Approach

I led the move in two stages. First a lift-and-shift to Google Cloud to get out of the old hosting. Then a re-platform until every workload ran as its own microservice on GKE, provisioned with Terraform.

Outcome

The domain runs fully cloud native across three regions, with each service scaling from dozens to hundreds of nodes as load requires.

Tools

• GKE
• Terraform
• Helm
• Microservices

The whole domain as code

Problem

Infrastructure for dozens of services across several regions needs to look the same everywhere, and every change needs an owner and a record.

Approach

Reusable Terraform modules define the building blocks. Terragrunt manages each environment on top of them. Helm charts describe every service, and Argo CD applies them, so Git is the only way a change reaches an environment.

Outcome

Every change to every environment is made and reviewed through Git. New environments and services come from the same modules instead of hand-built copies.

Tools

• Terraform
• Terragrunt
• Helm
• Argo CD
• GitOps

Cutting cloud cost by a third

Problem

A domain that scales to hundreds of nodes per service can spend a lot on Google Cloud, and the bill needs an owner.

Approach

I took on governance for the domain's infrastructure. Because everything is defined in Terraform and deployed through Argo CD, a cost change goes through the same pull request flow as any other change, so it is reviewed, rolled out gradually and can be reverted.

Outcome

Google Cloud costs for the domain fell by 34% in 2026.

Tools

• FinOps
• Governance
• GCP

Olinqua, Head of DevOps and Cloud Engineering (contract), Jan 2023 - Jun 2026

Delivery and Azure foundations for healthcare products

Problem

Healthcare products had no automated builds or deployments, and the Azure virtual machine estate was set up by hand for every new partner and customer.

Approach

I built CI/CD pipelines for the products and moved provisioning for new partners and customers to infrastructure as code, reviewed like application code. I also designed an Azure Landing Zones architecture as the standard foundation for networking, identity, governance and security.

Outcome

New partners and customers come from repeatable, reviewed code instead of manual setup. Azure costs fell after optimizing the virtual machine infrastructure, and most day-to-day operations are automated.

Tools

• Azure
• Landing Zones
• Infrastructure as code
• CI/CD