Surprisingly recent example that is really borderline[1], but they are indeed exceedingly rare (however I do count corporate/private RnD as being staffed by people coming from academia here).

1: https://www.quantamagazine.org/statistician-proves-gaussian-...