About
Senior platform engineer with 7+ years building internal platforms on AWS and Kubernetes. My customers are other engineers: I build the golden paths they deploy on — Crossplane and Terraform behind a simple API, reusable CI templates, base Helm charts, ArgoCD and FluxCD for delivery, and ephemeral preview environments on every pull request — then keep the whole thing cheap to run (30% off an EKS bill) and boring to operate. Never losing sense of humor.
Former professional opera singer, Cavalry Scout — United States Army, martial arts teacher, personal trainer, restaurateur, bodyguard, and volunteer animal rescuer. Allergic to stagnation and mediocrity, but lover of all things with 4 paws. Always hungry to grow.
Selected work
Fifty-four Terraform modules to Crossplane, in place →
A whole estate — 54 modules, roughly 190 distinct AWS resource types — moved onto a Kubernetes-native control plane without recreating the infrastructure underneath it. The interesting engineering is all in the handover procedure.
Taking 30% off an EKS bill without a capacity freeze →
Karpenter, honest requests, and preview environments that clean up after themselves. The savings came from consolidation and lifecycle, not from telling teams to use less.
Self-managed GitLab, assembled by a Crossplane control plane →
A full GitLab platform built from an empty AWS account — EKS, Aurora, ElastiCache, Cilium Gateway API, private PKI and a bridge into the corporate network — delivered as nine composable Crossplane function steps instead of a pile of hand-wired modules.
Where I've been
Built a multi-agent Claude Code / Codex workflow that works infrastructure tickets end to end — read-only AWS and EKS discovery, the fix, then drafted PRs, Jira updates and change requests behind a human approval gate — now a shared team tool. Split an over-trusted CI/CD role into separate plan and apply roles, shipped Terraform changes across a 12-account, two-organization AWS estate through PR pipelines, and mapped ownership of 38 legacy EKS services to unblock decommissioning.
Designed a multi-account Crossplane control plane and authored nine function-pythonic composition steps that stand up a complete self-managed GitLab platform from an empty AWS GovCloud account: EKS, Aurora PostgreSQL, ElastiCache Redis, S3, Pod Identity, Cilium Gateway API, cross-account private CA, Route 53, and an OKTA-backed NLB bridge into the corporate network. Added Chainsaw end-to-end tests to the composition's CI pipeline.
Brought Terraform and Terragrunt under GitOps with Burrito, moving deployments off developer workstations into one controlled place. Stood up automated drift detection to keep state continuously reconciled.
Platform team for MDE, the shared internal platform DHCS's modernization product teams deploy on. Replaced Terraform-automated infrastructure with versioned Crossplane configuration packages for a Kubernetes-native control plane. Moved push-based deployments to pull-based GitOps with ArgoCD and a shared product-environment Helm chart, and secrets from AWS Secrets Manager to HashiCorp Vault. Led the move to ephemeral environments. Built Datadog dashboards and alerting for a corporate Jira platform.
Built the company's entire platform on Amazon EKS with modular Terraform kept DRY via Terragrunt. FluxCD and ArgoCD app-of-apps, extensive Helm charts driving automated isolated preview environments, canary and blue/green with Argo Rollouts and Image Updater. Cut EKS spend by 30%.
Ran EKS applications and infrastructure in a high-volume environment with FluxCD, ArgoCD and Helm. Designed and adopted Karpenter for node autoscaling. Added SOPS to the release cycle and envelope encryption across the EKS environment. Built a cross-functional incident-response system spanning engineering, product and legal, and mentored junior engineers.
Built and managed entire AWS environments with modular Terragrunt-style Terraform, plus the Kubernetes clusters and GitOps CI/CD inside them. Authored Jenkins and GitLab pipelines and automated operational work in Bash and Python.
DevOps Engineer at MiTek Industries (2019–2020); Systems & DevOps Administrator at Blue Spruce Capital (2018–2019); Systems Engineer at Integritechs (2017–2018); and progressive IT and systems roles across Travelopia, California Resources Corp. and Gold's Gym from 2013.
What I go deep on
- Kubernetes EKS in production since 2020. Karpenter for node lifecycle, Cilium and Gateway API for networking, Helm for packaging, Argo Rollouts for canary and blue/green.
- GitOps ArgoCD and FluxCD app-of-apps at scale, Argo Image Updater, and Burrito to put Terraform under the same pull-based model as everything else.
- Infrastructure as code Terraform and Terragrunt kept DRY across accounts; Crossplane compositions when the control plane belongs inside the cluster. I have migrated production between the two in both directions.
- AWS Commercial and GovCloud, multi-account, Pod Identity over static credentials.
- Developer experience Golden paths over tickets: reusable CI templates (GitHub Actions, GitLab CI, CircleCI, Argo Workflows), base Helm charts, Crossplane APIs that hide the cloud, and preview environments that clean up after themselves.
- AI-assisted automation Claude Code and Codex agent workflows with MCP integrations, read-only guardrails and human approval gates — agents that do the toil, not the deciding.
- Secrets & access HashiCorp Vault, External Secrets, SOPS, Pod Identity over static keys, OKTA SSO — the guardrails built in, so teams never have to think about them.
- Observability & on-call Datadog, Grafana, Splunk, New Relic — and an incident-response process that engineering, product and legal all actually followed.