Welcome
Hi, I'm Hunter Wilson — a NOC Engineer who designs, builds, and operates production network and cloud infrastructure end-to-end.
What I Do
I'm a NOC Engineer at NetActuate, managing enterprise network infrastructure and BGP routing across a global tier-1 network spanning 40+ Points of Presence. My day job is traffic engineering, peering, and operations at scale; my nights and weekends are spent running my own hosting platform out of a colocation cabinet.
My focus is the full stack under a service: the physical servers, the hypervisor and network fabric, the automation that ties them together, and the customer-facing application on top. I hold CCNA, Network+, JNCIA, and FCA certifications and am pursuing RHCSA to round out the systems engineering side.
Featured Project: Piedmont Hosting
Production Cloud Hosting Platform
VPS, game servers, and web hosting — built from scratch on owner-operated colocated hardware
Piedmont Hosting is a full production cloud platform I designed, built, and operate solo. Everything from the physical servers in the colocation cabinet up through the customer-facing portal, billing, and REST API is my own work. It sells three products: metered per-resource VPS, fixed-tier game servers across 8 titles, and shared web hosting with email.
- Customer Portal: Next.js 14 App Router + Prisma + PostgreSQL, serving multiple domains via middleware-based host routing. NextAuth with TOTP 2FA and recovery codes.
- Billing: Live Stripe integration with metered subscriptions (per-core CPU, per-GB RAM, per-GB disk), fixed tiers, promo codes, automated dunning, and a $10 signup credit via Stripe setup mode.
- VPS Provisioning: Customer checkout fires a Stripe webhook that SSHes to the R730 hypervisor and invokes
mkvm— a custom KVM + cloud-init pipeline supporting 9 Linux distros, IPv4+IPv6 dual-stack, bandwidth QoS viatc, nftables-based cloud firewalls, snapshots, and full VM backups viavirtnbdbackup. - VPC Private Networking: Customers create isolated private networks (10.68.X.0/24) backed by Linux bridges on the hypervisor, with hot-attach/detach of NICs on running VMs.
- App Marketplace: 24 one-click Docker Compose apps (WordPress, Nextcloud, Immich, Vaultwarden, Grafana, Gitea, Jellyfin, Plausible, and more) deployed over SSH with a local Docker Hub pull-through cache.
- Game Servers: Pterodactyl API integration for 8 games (Minecraft Java/Bedrock, Rust, Valheim, ARK, CS2, Palworld, Hytale), free custom subdomains provisioned via the Cloudflare API.
- Web Hosting: CloudPanel-backed, supports WordPress, PHP, Node.js, Python, and static sites with automated SSL via Let's Encrypt; mailcow for email with DKIM/SPF/DMARC.
- REST API: Public
/api/v1with Bearer-token auth covering servers, firewalls, snapshots, and SSH keys — ~95 endpoints total across the portal and public API. - Monitoring: Prometheus + Grafana with a libvirt exporter for per-VM CPU/RAM/disk/network, internal and external Uptime Kuma, push alerts to a self-hosted ntfy server, and a Blackbox probe fleet for end-to-end latency.
- Edge: Cloudflare Tunnel fronting the portal so there are no open inbound ports for customer-facing services.
Other Projects
High Heat Forge — E-Commerce Platform
A custom Django 5.2 e-commerce platform I built from scratch for my father's bladesmithing and knife-making instruction business in Durham, NC. PostgreSQL, live Stripe checkout with server-side order finalization via webhooks, atomic inventory, Calendly integration for class scheduling, transactional email through the Piedmont mailcow, and a dark Tailwind UI. The site itself is hosted on my Piedmont Hosting platform — one of the first real customers on the web hosting product.
What Drives Me
I care about systems that just work — the ones that take real payments, serve real traffic, and don't need hand-holding because they were designed properly from day one. Running Piedmont solo means every layer has to be solid: if the firewall rules break, the automation stalls, or the portal throws a 500, there's no one else to fix it.
That pressure has pushed me to care deeply about documentation, monitoring, and automation. I'd rather spend a day writing a provisioning script than ten minutes doing it by hand, and I'd rather know a service is degraded from a Grafana alert than from a customer email.