Browse by topic
Nine curated clusters over everything I've written — each hub collects the posts on one practice area, newest first.
Handing an AI agent real access to your systems is an infrastructure decision, not a productivity hack.
11 posts
Running LLMs is an infrastructure problem before it is a modeling one.
18 posts
Kubernetes, containers, and the judgment calls around them — including when Docker Compose is the better answer.
27 posts
Infrastructure as code is the discipline that makes everything else repeatable.
26 posts
You cannot run what you cannot see — and observability bills scale faster than traffic.
21 posts
Security posts from the infrastructure side: zero-trust networking, WireGuard down to the hardware, securing MCP servers with OAuth, and keeping secrets away from AI coding assistants.
16 posts
Postgres is my default answer to most storage questions, and these posts explain why.
14 posts
Running compute at the edge, mostly on Cloudflare Workers — including the site you are reading.
15 posts
The layer under the platform: debugging latent kernel bugs in production, io_uring, eBPF, systemd lifecycle automation, and what Rust in the kernel means for operators.
10 posts