Berlin (CET) · Remote · Authorized to work in Germany · 4 weeks’ notice

Production Kubernetes on AWS, GCP and Azure

Eight years running production platforms, on fourteen years of infrastructure operations. The numbers below are the ones I was accountable for.

Hiring for

Evidence

Cut AWS EC2 spend by 30% with Karpenter node autoscaling, a move onto spot and reserved capacity, and automated off-hours shutdown of non-production environments.
Cut EC2 instance start-up from 14 minutes to 2 by pre-baking images, and application build time from 30 minutes to 10 with a shared dependency cache.
Led the migration of the main infrastructure codebase off generated CloudFormation onto Terraform, eliminating the configuration drift that came with it.
Own the EKS platform behind all production, staging and test environments — Helm-managed deployments, Istio service mesh, Karpenter provisioning, PostgreSQL and Kafka for event-driven services.
Built the monitoring and logging stack: Prometheus with Thanos for long-term retention, Grafana dashboards, and Loki for centralised logging.
Put Terraform plan and apply inside pull requests with Atlantis on a client platform, so product developers could change infrastructure behind review instead of filing a ticket.

30%

lower AWS EC2 spend

Karpenter, spot and reserved capacity, off-hours shutdown

14 → 2

minutes to EC2 instance start-up

pre-baked images; builds 30 → 10 min

367%

company growth in 2018

core infrastructure re-engineered ahead of Black Friday; no major incidents

The full timeline, the stack and what I am building now are on thefront page.