Josphat M.Private profile: Upwork shows its earnings only to signed-in members, so we check it every few weeks, not daily. MRR here is a rough estimate (Upwork's six-month earnings ÷ months active). Last read Nov 5, 2026.

Your Kubernetes cluster should be boring. Most of the ones I get called into are not: no infrastructure as code, one person who knows the deploy steps, secrets sitting in git, and a cloud bill nobody can explain. I fix that, then hand back infrastructure your own team can run without me. Nine years building and running production infrastructure. Right now I am a senior platform engineer running Kubernetes on GCP and AWS. Before that I ran the EKS estate for a fintech, hundreds of microservices across a lot of environments, and built OpenShift clusters on bare metal at East Africa's largest telco, which is where you learn what the managed control plane had been doing for you. WHAT CLIENTS HIRE ME FOR - Kubernetes done properly, the first time. EKS, GKE, AKS or self-hosted on your own hardware. Autoscaling that scales on something real, network policies, ingress with certificates that renew themselves, pod disruption budgets, and a cluster that comes up from a cold start in the correct order rather than in whatever order the manifests happened to apply. - GitOps you can hand to a product team. Argo CD with per-environment overlays off one set of Helm charts, sync waves so dependencies land before the workloads that need them, automatic rollback on a failed sync, and single sign-on instead of a shared Argo CD password. Environments differ by configuration, not by a forked copy of the repository. On the platform I built, twenty-two services across five teams onboard by writing a single values file against one shared chart. - Everything as code. Terraform, OpenTofu and Terragrunt with reusable modules, the plan running on the pull request and the apply on merge, so infrastructure review looks like code review and nobody applies from a laptop. I have built a framework of more than twenty-five reusable modules with Atlantis, and consolidated two dozen scattered Terraform repositories into one governed monorepo with a migration designed so no state ever had to move. - Cloud bills. Most are not expensive because of one big mistake but because of forty small ones nobody owns: idle non-production, oversized nodes, Kubernetes requests set to a guess, overlapping services, and tagging too loose to say which team spent what. I audit read-only first and change nothing, then give you every finding with the saving, the risk and the effort attached. The reason most cost reports never get implemented is that whoever wrote them cannot say which changes are safe. I can, because I would be the one on call afterwards. - Migrations and rescues. On-prem to cloud, VMware or Proxmox to Kubernetes, legacy applications into containers, and the less pleasant kind of work where production is down, there is no documentation, and the engineer who built it has left. Send me access and the error and I will tell you what I find before I bill you for fixing it. - Secrets, access and backups. Vault and External Secrets, single sign-on, least-privilege IAM, no long-lived cloud keys anywhere, and backups I prove by restoring them instead of by showing you the schedule. HOW I WORK A short call to find out what is actually broken, then a written plan with the scope, the sequence and the cost before I start. Daily written updates in your channel, so you never have to ask where things are. You get the code, the runbooks and a walkthrough, which is the difference between infrastructure a team can run and infrastructure that walks out the door when the contract ends. BACKGROUND CKA, AWS Certified Solutions Architect, AWS Certified SysOps Administrator, and Red Hat Certified Engineer. I also write the Linux and DevOps guides at computingforgeeks com, which is where a lot of engineers land when they paste an error message into a search engine, and I test everything on a multi-node Proxmox and Kubernetes lab before it goes anywhere near production or into an article. CORE STACK Kubernetes (EKS, GKE, AKS, OpenShift, kubeadm), Helm, Kustomize, Argo CD, Argo Rollouts, Crossplane, KEDA. Terraform, OpenTofu, Terragrunt, Atlantis, Ansible, Packer. AWS and GCP, Cloudflare. GitHub Actions, GitLab CI, Jenkins. Prometheus, Grafana, Datadog, Loki, Sentry. Vault, External Secrets, cert-manager, OIDC and SSO. PostgreSQL, MySQL and Galera, MongoDB, Redis, Elasticsearch. Linux (RHEL,Ubuntu, Debian), Proxmox, VMware, KVM, Ceph. Python, Bash, Node.js and TypeScript. Tell me what is breaking or what you are trying to ship, not just the job spec. If I am not the right person for it I will tell you that and point you at who is.