Skip to content

Now / November 14, 2026

Building the bridge from product ownership to platform ownership.

The production work, structured learning program, and practical evidence I’m using to make the transition credible.

Current focus

Kubernetes and traffic governance, running on a three-node GCP cluster bootstrapped with kubeadm. An ALB fronts the cluster, Helm 4 installed Traefik, and this site runs on it. CKA is at the end of November.

Working

I’m currently a full-stack and platform infrastructure engineer at Molar Intelligence in Hangzhou, China. I’m the primary engineer across several public products and an internal publishing platform, working from public interfaces and content APIs through storage, publishing workflows, observability, and production delivery.

Owning these systems after launch—especially evolving Docs Hub into a self-service platform used by seven colleagues—clarified the direction I want to pursue: deeper responsibility for the platforms, delivery systems, and operational foundations that product teams depend on.

The cluster this site runs on

I now run a production DIY Kubernetes cluster on GCP with three nodes: one control-plane node and two worker nodes. I bootstrapped it with kubeadm, put an ALB in front of it, installed Helm 4, and used Helm to deploy Traefik as the ingress controller.

This website is hosted inside that cluster. It is the first real workload, and I’ll keep deploying my personal applications there.

The upskilling program

On August 28, 2026, I began a structured platform-engineering program built around one principle: learn fewer technologies, but use each one to produce credible operational evidence.

The current stage combines:

  • the complete KodeKloud CKA Learning Path and hands-on Kubernetes labs;
  • building and operating the GCP cluster;
  • a production-inspired entry-traffic governance case study;
  • a smaller, parallel system-design track;
  • a target cadence of 22 hours per week, split roughly 80% toward Kubernetes and the case study and 20% toward system design.

I still use kind for quick local experiments, but the production path is now the GCP cluster.

Current platform case study

The practical problem is the full request path:

DNS → ALB → Traefik ingress → Service → Pod

I’ve implemented this path on the cluster: kubeadm bootstrapped the nodes, Helm 4 installed Traefik, and the ALB fronts the ingress layer. I’m documenting the topology, clarifying component ownership, comparing routing designs, and building a searchable route-to-backend inventory. The final case study will also include failure drills, entry-layer observability, a troubleshooting runbook, migration and rollback decisions, and an anonymized working implementation.

The goal is not to collect Kubernetes commands. It is to show that I can understand, redesign, operate, and explain a production traffic path.

Certification roadmap

CKA — end of November 2026

I’m taking the Certified Kubernetes Administrator exam at the end of November 2026. The cluster work, KodeKloud labs, and troubleshooting practice are the preparation; the exam is a checkpoint for practical Kubernetes fluency, not a substitute for it.

CNPE — target window: April–July 2027

The Cloud Native Platform Engineer certification comes later, after the wider platform project is substantially complete and has been tested through deployment, security, observability, GitOps, progressive delivery, backup and recovery, and failure drills.

I’ll attempt CNPE when simulator results show exam-speed gaps rather than major conceptual gaps. It should validate the platform work, not be the source of the knowledge.

Building alongside the program

This site now runs inside the cluster, and I’ll deploy more personal applications there as they’re ready. I’m also continuing to develop Iokeep v3, my long-running note-taking application, as another workload to package, deploy, observe, and operate.

Longer-term direction

The immediate target is a platform, cloud-platform, or SRE-oriented engineering role. Longer term, I want to grow into AI platform engineering: building, securing, observing, and operating the systems that AI applications run on while retaining the application-level judgment that comes from full-stack work.

Updated as the work—and the evidence—changes.