DevOps · Cloud · Platform Engineering

Kyle Mapue

DevOps / Research Engineer

Brisbane, Australia

I build and operate reliable cloud infrastructure, designing Kubernetes platforms, GitOps workflows and CI/CD pipelines that ship software safely and at scale.

  • Kubernetes
  • GitOps
  • OpenStack
  • CI/CD
  • Terraform
Currently: Research Engineer @ University of Queensland
scroll ↓

Selected work

Platforms I've built & operated

A few projects that took me from theory to running, observable systems.

Neurodesk: Open-Source Platform Engineering

Open source · org admin

I'm organisation administrator for Neurodesk, an open-source neuroimaging platform used by researchers worldwide. Across five repositories I own its cloud and CI/CD infrastructure end to end, and contribute to the science notebooks themselves. 40+ merged pull requests and counting; co-author on a 2025 OSF preprint.

40+
PRs merged
16+
VMs to ephemeral pods
20-30%
faster CI/CD
  • ARC Runners
  • Terraform
  • GitHub Actions
  • GHCR / Quay
  • Apptainer
  • Multi-AZ
Selected contributions
  • Re-architected CI from 16+ always-on VMs to Terraform-provisioned Kubernetes ARC ephemeral runners, with a second multi-AZ cluster for redundancy (~99% uptime).
  • Built the per-pull-request preview-deploy system for neurodesk.org: bot-comment live previews, safe even for external contributors, with gated staging to production promotion.
  • Migrated container builds onto self-hosted ARC pools and moved production images to GHCR with a Quay mirror and native arm64 multi-arch builds.
  • Created a lightweight, workflow-based review system for community-contributed teaching notebooks (NeurodeskEDU): JOSE-inspired but pure pipelines, no bot or VM, with review badges and UUIDs.
  • Implemented secrets management and least-privilege DevSecOps controls on the self-hosted ARC runners, hardening the CI supply chain.
  • Currently building content-addressed distribution of Apptainer .sif images as OCI artifacts via ORAS, pulled by digest from a registry instead of hosted as standalone files.

Todo Platform: Cloud-Native on GKE

Cloud-native

An end-to-end GitOps platform for a multi-service Todo app on Google Kubernetes Engine: Kustomize overlays for staging/production, GitHub Actions CI/CD, and Argo Rollouts canary deploys with metric-driven automatic rollback.

3
microservices
2
GitOps environments
Auto
canary rollback
  • GKE
  • Kustomize
  • Argo Rollouts
  • NATS
  • Prometheus
  • Grafana

Log-Pong: kubectl to GitOps

A stateful two-tier app (Node.js/Koa + PostgreSQL) taken from manual kubectl deploys to a full ArgoCD GitOps workflow with readiness probes and progressive canary releases for zero-downtime delivery.

0
downtime deploys
2
service tiers
GitOps
single source of truth
  • Kubernetes
  • ArgoCD
  • GitHub Actions
  • Docker

mapuekyle.com: This Site

Personal

My portfolio, rebuilt with Astro and an interactive three.js infrastructure scene. Previously provisioned on a DigitalOcean droplet with Terraform, Ansible and GitHub Actions. The story of both builds lives in the journal.

  • Astro
  • three.js
  • Netlify
  • Terraform
  • Ansible

Experience

Where I've worked

  1. Research Engineer (DevOps)

    Jun 2025 - Present

    The University of Queensland · Brisbane, AU

    • Organisation administrator for the open-source Neurodesk research platform, responsible for the quality, security and scalability of its cloud and CI/CD infrastructure.
    • Re-architected the platform from 16+ always-on VMs to a Terraform-provisioned Kubernetes ARC setup with ephemeral, scale-to-zero runners, migrating one repository at a time with zero hard cutover, then standing up a second cluster in another availability zone from the same code for active redundancy.
    • Designed a per-pull-request preview deployment system for neurodesk.org (a composite of 4 GitHub Pages repos), including sandboxed, secret-scoped previews safe for external contributors, paired with gated staging to production releases.
    • Standardised reusable YAML pipeline templates and led a secrets audit / least-privilege uplift across the org’s repositories (neurocontainers, neurodeskedu, neurodesktop), with documentation and team knowledge transfer.
    • Built a lightweight, workflow-based review system for community-contributed teaching notebooks (NeurodeskEDU), moved production container images to GHCR with a Quay mirror and arm64 multi-arch builds, and am developing ORAS-based content-addressed distribution of Apptainer .sif images.
    • Operate a Magnum-provisioned OpenStack/Kubernetes cluster with a Helm-managed observability stack (Prometheus, Grafana, Loki); designing ARM-based builders for multi-architecture container workflows.
    16+
    always-on VMs retired
    20-30%
    faster CI/CD throughput
    ~99%
    platform availability
    • Kubernetes
    • Terraform
    • Helm
    • GitHub Actions
    • ARC
    • OpenStack Magnum
  2. DevOps Engineer

    Jan 2023 - May 2025

    Globe Telecom · Metro Manila, PH

    • Part of the pioneer team that automated critical department-wide tasks, significantly improving company-wide efficiency and cost savings.
    • Delivered solutions across varied use cases with the goal of automating network alarms end-to-end.
    • AWS
    • Kubernetes
    • Linux
    • Python
    • Drools
  3. Frontend Web Developer

    Aug 2021 - Feb 2022

    Fourello · Manila, PH

    • Designed and built core parts of a web application, covering both the interface and the user-facing behaviour.
    • Vue.js
    • JavaScript
    • Node.js
    • Git
  4. IT Intern

    Feb 2022 - May 2022

    Linde Philippines, Inc. · Manila, PH

    • Supported system administration, resolved tickets and handled daily operations.
    • Oversaw deployment and setup of new devices for distribution.

About

Hi, I'm Kyle 👋

I'm a DevOps and platform engineer with 5+ years across the full delivery lifecycle. I started in frontend, moved into automating network operations at Globe, and now design and run Kubernetes/OpenStack build infrastructure at the University of Queensland, where I'm also organisation administrator for the open-source Neurodesk research platform.

I learn by building real, end-to-end platforms, and writing honestly about the parts that broke along the way. When something fails at 2am, that's usually where the best lessons are.

Education B.S. Information Technology University of Santo Tomas Cum Laude · 2018 - 2022
Toolbox

Orchestration & IaC

  • Kubernetes
  • Docker
  • Terraform
  • Ansible
  • ArgoCD
  • Argo Rollouts
  • Kustomize
  • Helm

CI/CD & GitOps

  • GitHub Actions
  • Actions Runner Controller
  • GitOps
  • GitHub Pages

Cloud Platforms

  • AWS
  • Azure
  • Google Cloud
  • OpenStack (Magnum)

Observability

  • Prometheus
  • Grafana
  • Loki

Languages

  • Python
  • JavaScript / Node.js
  • PHP
  • Vue.js
  • React
  • SQL
Certifications
  • AZ-400: Implementing DevOps Solutions Microsoft · 2023
  • SC-900: Security, Compliance & Identity Fundamentals Microsoft · 2022
  • AZ-900: Azure Fundamentals Microsoft · 2022
  • Certified Cloud Practitioner AWS · 2022
  • IT Automation with Python Google / Coursera · 2020
  • IT Passport (PHILNITS) ITPEC · 2021

Publications

Research & writing

Get in touch

Let's build something reliable.

I'm happy where I am at UQ, but always open to interesting infrastructure problems and the right opportunity. The fastest way to reach me is email.