Platform engineer, four years in. The systems below are mine to keep standing, and to make boring enough that nobody thinks about them.
The EKS platform product teams ship to. I set the deployment patterns everyone uses: reusable Helm configurations covering resource limits, health checks and rolling updates, now adopted across four product engineering teams. The goal is that shipping a service requires knowing nothing about how any of it works.
How code gets from merged to running. I moved customer-facing workloads off manual kubectl releases onto GitOps with Argo CD, so every change is declarative, peer-reviewed and revertible.
How services prove who they are. I removed long-lived AWS keys from application and deployment workflows in favour of identity-based authentication, and took plaintext database credentials out of the deployment path entirely. Fewer secrets in existence means fewer secrets to leak.
I’m on the hook when the platform misbehaves. I built an AI-assisted triage workflow that correlates monitoring alerts, Kubernetes events, application logs and recent deploys into a single picture. Most of the saving is time nobody spends assembling context by hand.
I believe in learning for life, not just for the next job. That is most of why I ended up in infrastructure: the ground keeps moving, and you are never quite finished.
The other half of it is people. I am on the core team of the AWS Student Builder Group, and I help grow a college club that is now around 35 members and still adding. As a club we take part in AWS Jam Showdowns: immersive team challenges where you solve real problems in a live cloud environment, built for complete beginners and tinkerers alike.
Teaching is the fastest way I have found to spot the gaps in what I know. And almost everything I understand today came from someone who took the time to explain it to me, so passing that on feels like the right way to say thank you.
Side projects, homelab experiments, and whatever else survives the weekend. Still collecting the good ones.