DevOps Engineer — Kubernetes, GitOps & release automation
The knowledge to answer 'which services are exposed to the internet, and who owns them?' already exists — it's just scattered across a half-dozen tools that don't talk to each other. A knowledge graph is how you make it queryable.
A walkthrough of platform-kg: extract facts from ArgoCD, Kubernetes, Helm, GitHub, Terraform, Vault, and Istio, turn them into RDF triples, and load them into GraphDB — a graph of your actual cluster you can query in a browser.
Once you have a graph of your platform, the obvious question is: now what? It turns out there's a ladder of answers — inventory, ownership, dependencies, blast radius — each rung built on the one below, with real SPARQL at every step.
In 2026, the hard part of putting an AI agent on your infrastructure is no longer the AI — it's giving the agent something trustworthy to reason over. A platform knowledge graph is exactly that context layer, wired in over MCP.
Full-stack Helm-chart generation platform in Go & React. One declarative ChartpressConfig becomes a complete umbrella chart with per-service Deployment/StatefulSet/DaemonSet subcharts — exposed as a Cobra CLI, an HTTP API, and a React/Vite web app.
Cookiecutter generator for Helm-focused Argo CD GitOps repos — scaffolds ApplicationSets, bootstrap manifests, and per-cluster Helm values across dev/stage/prod with CI-driven progressive promotion.