Blog
Notes from building and operating the Home AI Lab - covering local LLM inference, network automation, identity, and self-hosted agents. Earlier posts include August and October 2026 update blocks at the bottom as the lab evolved.
Resurrecting Google Glass Enterprise Edition 2 with a Local Voice Gateway (And Four Ghost Bugs Along the Way)
Turning a retired Android 8.1 smart-glasses headset into an ambient lab terminal backed by a dedicated Whisper + Coqui TTS VM, plus four infrastructure mysteries where every health check returned 200 OK while the system was broken.
Blackwell FP4, the Thinking-Token Paradox, and Splitting MCP into Read and Write
Retiring Gemma 4 for Swift-Qwen3.8-27B on native Blackwell NVFP4 tensor cores, discovering why Qwen's thinking mode behaves the exact opposite of Gemma's on tool calls, and splitting five MCP servers into read-only and admin profiles.
The Backup That Died Silently, and the Mirror That Worked Too Well
A two-act homelab post-mortem: discovering that two independent backup legs had been failing in silence, and then discovering that the replacement was syncing unencrypted config dumps to cloud storage.
Why I Took the LLM Out of My AI Assistant's Morning Report (And Taught It to Sound Like Homer Simpson)
What happens when an autonomous cron agent quietly stops calling its tools for eight days, why deterministic scripts beat LLM health checks, and how a two-stage Piper + RVC daemon delivers voice notes in under ten seconds.
sequin.au: Writing the Lab into Existence
On building a blog to document a home lab, why documentation is infrastructure, and what happens when the project that chronicles everything becomes the thing that everything was for.
Building the Home AI Lab
An overview of REX - the Windows 11 host machine running a fleet of Hyper-V VMs, local LLMs, and MCP-connected network infrastructure.
Spec-Driven Development: How to Stop AI Agents from Rewriting Everything
Adopting a constitution-first development methodology to prevent AI coding assistants from being creative in ways you did not request. Also: LangGraph for autonomous agents.
Clawdia: The AI Assistant That Texts You Security Reports at Breakfast
Building an AI home assistant on UbuntuClaw with local LLM inference, Piper TTS, Whisper ASR, and WhatsApp integration. Because your network anomalies deserve a Welsh accent.
MCP: Giving AI Agents the Keys to the (Carefully Segmented) Kingdom
Building four custom MCP servers so Claude can query your Juniper switch, check OPNsense firewall rules, monitor Aruba APs, and manage the Synology NAS. What could go wrong.
LiteLLM: One Endpoint to Rule Them All
Deploying a LiteLLM proxy on GCP to unify Claude, Gemini, and local Gemma4 behind a single OpenAI-compatible API. Also: why nginx and Redis are doing more work than you'd expect.
Gemma4 at Home: Running 26 Billion Parameters on Consumer Hardware
vLLM, WSL2, an RTX 5090, and the specific joy of watching a 26B model load into VRAM without catching fire.
Vulnerability Management for People Who Enjoy Finding Their Own Problems
Running Greenbone/GVM and daily Nmap discovery on KaliLinux. Because knowing what's on your network is better than finding out via a breach notification.
Trust No One: Building an Internal PKI, Identity Platform, and Secrets Vault
Self-signed certificates are the 'password123' of TLS. Here's how DebIdentity got its own Certificate Authority, an enterprise identity provider, and a password vault - all in Docker.
Six VLANs and a Juniper Switch: Networking the Lab Properly
Flat networks are fine until they aren't. Segmenting a home lab with enterprise gear, OPNsense, and the kind of routing table that impresses nobody at dinner parties.
REX: Or How I Justified Buying a 7950X3D for 'the Lab'
Every home lab starts with a lie you tell yourself. Mine was: 'I just need one decent machine.'