// hey, I'm

Thad Roebuck

platform engineer.

My passion is building the self-service tooling and infrastructure that delights developers by enabling them to ship excellent products faster, safer, and better.

GitHub ↗ LinkedIn ↗ Get in touch

01. about

I'm a platform engineer who brings a product mindset to infrastructure. My job isn't just to run clusters — it's to build the self-service platforms, paved roads, and developer tooling that other engineers adopt to deliver excellent products.

I believe in GitOps as the operating model, not just a deployment strategy. Declarative everything. Pull-based reconciliation. Git as the single source of truth. Over time, this model has gotten so good that it's boring, which is exactly how I like it.

Right now I'm focused on multi-tenant Kubernetes platforms, internal developer portals, and making the "paved road" so good that teams choose it over the shortcut.

stack

  • Kubernetes
  • Argo CD
  • Flux
  • Terraform
  • Helm
  • Prometheus
  • Grafana
  • Go
  • Python
  • Backstage
  • GitHub Actions
  • Crossplane
  • Nix
  • Linux

02. projects

Internal Developer Platform ↗

problem

Developers waited days for environment provisioning and had no self-service path to production.

solution

Built a Backstage-based IDP with custom plugins for service scaffolding, environment provisioning via Crossplane, and automated golden path templates.

impact

Reduced new service onboarding from 5 days to 15 minutes.

  • Backstage
  • Crossplane
  • Kubernetes
  • TypeScript
  • PostgreSQL

GitOps Fleet Management ↗

problem

Managing 20+ Kubernetes clusters across 3 cloud providers with manual kubectl deploys and config drift.

solution

Designed a multi-cluster GitOps architecture using Argo CD ApplicationSets with a monorepo strategy. Implemented progressive delivery with Argo Rollouts.

impact

Zero-touch deployments across all clusters. Deployment frequency increased 4x. Config drift incidents dropped to zero.

  • Argo CD
  • Argo Rollouts
  • Kustomize
  • Helm
  • Go

Observability Platform ↗

problem

Fragmented monitoring across teams, no standardized SLOs, alert fatigue from noisy pages.

solution

Built a centralized observability stack with Prometheus, Grafana, and Loki. Created Terraform modules for teams to declare SLOs as code with auto-generated dashboards and alerts.

impact

MTTR reduced by 60%. Alert noise dropped 80%. Every service ships with SLOs from day one.

  • Prometheus
  • Grafana
  • Loki
  • Terraform
  • Go
  • OpenTelemetry

03. experience

Jan 2025 — present

Systems Engineer (Sole DevOps Lead) @ Xylem Tree Experts

  • Engineered a strategic migration from Azure to AWS using Crossplane for IaC, cutting monthly cloud costs by 20%
  • Deployed and managed 4–6 EKS clusters hosting 24+ microservices each with CNPG, Envoy, and External-DNS
  • Accelerated deployment frequency from quarterly to bi-weekly (6x) by implementing ArgoCD and GitOps workflows
  • Owned Keycloak IAM and External-Secrets implementation across all environments
  • Built self-service internal developer platforms to standardize deployments and reduce developer friction
Oct 2022 — Oct 2024

Technical Support Specialist (Tier 2) @ Decisions LLC

  • Diagnosed and resolved critical software incidents for high-value clients, reducing MTTR by 15%
  • Managed Windows/Linux server environments and SQL database troubleshooting for performance bottlenecks
  • Automated server health checks and log analysis with PowerShell/Bash scripts
  • Authored container-related technical docs that reduced Tier 1 escalation rates by 35%

04. contact

I'm currently looking for my next role on a platform engineering team. If you're building developer infrastructure, internal platforms, or Kubernetes-native tooling, I'd love to talk.

$ send_message