Principal Platform / DevOps Engineer
Kubernetes · Cloud-agnostic infrastructure · GitOps · Developer Experience
I'm a principal-level platform engineer. I build the paved roads that let application teams ship safely and fast — internal developer platforms, self-service infrastructure, and the guardrails that make the right way also the easy way. My work lives where reliability, developer experience, and cloud economics meet.
My north star is leverage: turning a fragile, tribal-knowledge runbook into a single declarative API that anyone on the team can use correctly on their first try. I gravitate toward systems that are boring on purpose — idempotent, observable, and easy to reason about at 3 a.m. — because the flashy part of infrastructure is that nobody has to think about it.
I work cloud-agnostic by default: Kubernetes-native primitives and portable tooling over lock-in, so the platform outlives any one provider decision. And I hold my own code to the same bar I'd ask of a team — everything below is real, tested, and CI-gated, so click in and read it.
- 🧭 Platform engineering — internal developer platforms, golden paths, self-service APIs
- ☸️ Kubernetes — operators & controllers, multi-tenancy, policy-as-code, workload isolation
- 🏗️ Infrastructure as Code — Terraform modules, GitOps (Argo CD / Flux), immutable golden images
- 🔭 Reliability & observability — SLOs & error budgets, Prometheus/Grafana, actionable alerting
- 🔐 Secure by default — least-privilege RBAC, network policy, CIS hardening, supply-chain hygiene
"Make the right thing the easy thing, then get out of the team's way."
| Project | What it is | Stack |
|---|---|---|
| kubtenant | A Kubernetes operator that turns soft multi-tenancy into a single Tenant CRD — provisions guardrailed namespaces (quotas, limits, NetworkPolicies, owner RBAC), garbage-collected via owner references. Prometheus metrics, 80%+ controller coverage. |
Go · controller-runtime |
| terraform-kubernetes-platform-baseline | Cloud-agnostic Terraform module that bootstraps a production add-on baseline (ingress-nginx, cert-manager, external-dns, kube-prometheus-stack, metrics-server, Argo CD) onto any cluster. Each component independently toggleable. | Terraform · Helm |
| golden-image-factory | A multi-cloud golden-image pipeline — bakes hardened, CIS-flavored Ubuntu images for AWS, Azure & GCP from one source of truth, with Ansible provisioning and node_exporter baked in. | Packer · Ansible |
| ktx | A fast, delightful terminal UI for fuzzy-switching Kubernetes contexts and namespaces — plus scriptable subcommands. Pure kubeconfig edits, no cluster calls. | Go · Bubble Tea |
| sre-toolbox 🔗 live | A zero-backend web toolbox of everyday SRE calculators: error-budget/SLO, availability nines, latency percentiles, cron explainer, JWT decoder. 100% client-side. | Vanilla JS · GitHub Pages |
apiVersion: platform.hariboddapati.dev/v1alpha1
kind: Tenant
metadata: { name: acme }
spec:
namespaces: [acme-dev, acme-staging, acme-prod]
owners: [{ kind: Group, name: acme-platform-admins }]
resourceQuota: { cpu: "16", memory: 32Gi, pods: "80" }
networkIsolation: trueidempotent reconciliation · native GC, no finalizers · distroless non-root image · race-tested, CI-gated
Building reliable, self-service platforms — one paved road at a time.