Technical blog about Infrastructure, Kubernetes, AI, and more.
Using Claude Code skills to automate k3s diagnosis and repair on my homelab, plus LLM inference tuning that pushed Hermes TPS from 21 t/s to 32+ t/s.
How I put an ARC self-hosted runner inside my homelab K8s cluster so GitHub Actions workflows can reach internal services without tunnels.
Using LiteLLM Proxy to unify Lemonade (AMD ROCm llama.cpp) and a ChatGPT subscription behind a single OpenAI-compatible API in my homelab.
Commit message generation was quietly draining my Copilot free quota. Fixed by pointing the utility model at a local Gemma-4-E4B via LiteLLM.
Running llama.cpp on my k3s + AMD GPU cluster kept hitting memory access faults. The culprit: a bug in MES firmware 0x83 shipped with amdgpu-dkms-firmware.
I set up Strix Halo as a k3s worker via Incus VM + VFIO, then hit a wall: once the GPU enters a dirty state, recovery is impossible without bare metal.
device-plugin gives the GPU to one Pod at a time. Here's why I switched to DRA on k3s, and three Strix Halo-specific issues I had to patch around.
How I joined GMKtec EVO-X2 (Ryzen AI MAX+ 395) to my k3s cluster as a GPU node via Incus VFIO, covering APU-specific passthrough gotchas.
Migrating from MicroK8s to K3s. Real-world insights on infrastructure rebuilding, from an Ubuntu 26.04 twist to Kubeconfig traps and safe TLS switching.
How I built a Discord BOT in Go to securely interact with my private homelab server without exposing it to the internet.
How I built a self-healing, automated refactoring pipeline using Codex's subscription capacity and Temporal on a home Kubernetes cluster.
Implement popular post rankings on Astro static sites using GA4 data. Covers build-time fetching, content integrity checks, and CI fallback strategies.
