
Small local llm lab with an old mining GPU
Testing a cheap headless mining GPU as a local AI inference card for Ollama and homelab benchmarks.
A running log of the systems, networking, and hardware projects I'm tinkering with, plus notes, fixes, and the little victories that come with them.

Testing a cheap headless mining GPU as a local AI inference card for Ollama and homelab benchmarks.

Moving my workstation to Fedora Atomic with Sway for a boring, predictable, and fast daily workflow.

Inline WAF and SPOA routing with HAProxy to harden public-facing services and trim noisy traffic early.

Declaratively cloning VMs from templates, wiring Cloud-Init snippets, and pushing configs straight through the Proxmox API.
Deep dives, build logs, and quick references.
One Card AI Inference Lab with a Used Mining GPU !
Fedora Atomic Sway as a daily driver !
Why I Built This Flow
What is WireGuard?
OPNsense firewall inside my Proxmox, but my WAN is already behind a private IP. That means double NAT, and it changes a few defaults. This is the exact VM se...
Repositories
Installing IDS/IPS on pfSense with Suricata
Why I Needed This
Use cases
ZFS infosZFS (Zettabyte File System)