Engineers' work is also externally verifiable, e.g. by unit tests for software, but I'm assuming by other sorts of automated protocols for civil engineering. I would hope a bridge is not built without triple checking the various outcomes.