AI Infrastructure Platform
Self-hosted everything on a single server: 64 GB RAM, 11 TB, RTX 3050, 38 containers. Nextcloud, Matrix, my media stack, all behind Caddy with TLS, on a VLAN-segmented network.
Runs my local LLM inference via Ollama. Four more VMs for Home Assistant, GNS3, remote desktop, and dedicated agent workloads.