No, though at my current company we work under the assumption that our work will be translated at some point in the future. The main effects are (a) more emphasis on clear vocabulary choices in cases where native-English readers would understand how to resolve ambiguity (fine with me; I'm a fan of precision anyway) and (b) a push to simpler sentence structures. The latter has sometimes been challenging when the concept itself is complex; our editor (when we had an editor) would sometimes break long sentences into multiple sentences and lose the tight coupling between the clauses. So that's been a bit of a challenge for me.

But we have yet to see the results, to know if we're actually doing the right things to support good translation. We're following what we've learned about best practices, but our work hasn't been put to the test yet.

(Oh, one other effect I just thought of: labelled diagrams are tricky. I don't have tons of those, and right now I'm just blowing this off -- numbered callouts with separate text don't work well for some things.)