I use independent infrastructure projects to explore how professional operating models feel in practice. The point is not to pretend a lab is production. The point is to learn which controls make work clearer.
Controls worth keeping
- Bounded actions with clear intent.
- Readable state before and after a workflow.
- Approvals or pauses before destructive operations.
- Small evidence packs that explain what happened.
The balance
Too much process kills curiosity. Too little structure makes the work hard to share. A good hobby control plane sits in the middle: enough discipline to make the experiment understandable.