About

I'm a platform and reliability engineer with over a decade building and operating Kubernetes infrastructure — GCP, AWS, and Azure, fleets past 300 nodes. Most of that decade has been the same combination in different shapes: observability that engineers actually act on, cost and capacity work, and automation that replaces a process someone used to do by hand.

I'm currently a Staff Cloud Engineer at Redox, a healthcare data interoperability company, where I run the observability and cost-engineering side of a multi-cluster platform. The most recent chapter of that work — an AI-assisted engineering harness that started as a memory system and grew into autonomous pull-request review — is what most of the writing on this site is currently about.

Before Redox, I founded and ran Brokee, a DevOps skills-assessment startup: hands-on cloud labs instead of whiteboard interviews. I was the sole engineer on the first version — frontend, backend, infrastructure, all of it — then hired and grew a small team while continuing to set architecture. We ran 1,000+ engineers through assessments before the company wound down. Earlier roles — EclecticIQ, HAL24K, Blockport, KPN, Accenture — were variations on the same theme across cyber threat intelligence, AI decision-support, a cryptocurrency exchange, a national telecom, and consulting: Kubernetes, observability, and infrastructure-as-code, on whichever cloud the job required.

I contribute upstream where I run into a real bug rather than file it and move on — recent fixes and reports against OpenTelemetry eBPF Instrumentation, Cilium, KEDA, Kubespray, and Helmfile, among others.

Before any of this, an MSc thesis at SINTEF in Oslo turned into two published papers on model-driven multi-cloud deployment — cited about 125 times since.

This site is the canonical home for my technical writing. Find me elsewhere:

• GitHub
• LinkedIn
• Google Scholar
• RSS