Projects seen through the decisions, not the architecture: the problem they were meant to solve, what went wrong along the way, and what I would do differently today.
How I check that a backup can bring back both the data and a usable application
The nightly backup kept reporting success. Trying to get my data back exposed gaps those reports had never caught. I now test the saved files, the restored data, and whether the application can use it.
From “it probably works” to an alert on your phone — metrics, logs, and local AI that summarizes alerts in plain language
When an infrastructure is used daily by a whole family, an outage stops being merely technical — somebody misses a film, somebody cannot find a document. A case study on what you have to measure to find out first, and why that differs from watching graphs. The same principle carries over to a small business.
How a shared shopping list turned a homelab into a small production system
Home App began as a shared shopping list and grew into a private home system: tasks, calendar, notifications, notes, Home Assistant, and a local AI agent. The best production test turned out to be the family itself.