Writing
Notes on platforms, Kubernetes, observability, CI/CD, and hiring engineers now that AI is part of the job. Written from work I have done.
Hiring engineers when AI is in the room
Banning AI from technical interviews tests a job that no longer exists. What I look for instead, after thousands of interviews, and why the process is the signal now.
How I build with AI coding agents every day
A DevOps engineer's working notes on using Claude Code and agent tooling for real infrastructure work: what to hand over, what to keep, and the guardrails that matter.
The first 30 days with an unfamiliar infrastructure
How I approach a platform I did not build: map it before touching it, find what actually pages people, and earn the right to change things with small wins.
Broken on purpose: designing interview challenges that feel like the job
How we build EasyEnv challenges as Ansible recipes that break a real machine in a believable way, and what makes a good one.
Kubernetes at the edge, on x86 and ARM
Lessons from running a 100+ node Kubernetes cluster that spanned cloud servers, NVIDIA Jetsons and Raspberry Pis: multi-arch images, flaky links and installs as code.
OpenTelemetry before the vendor
Why I instrument with OpenTelemetry first and pick a backend second, from building an internal developer platform inside a large bank.
One CI/CD for 200 projects
Unifying build and deploy across 200+ projects at a payments company: a single SCM strategy, a Jenkins nobody has to worry about, and caching the outside world.