[compsciphd]

when I was a phd student at Columbia, we had a student in our lab who was a full time employee at TJ Watson and his work at Watson was part and parcel of his phd. Basically supervised by both his IBM manager and my advisor.

When I was looking at programs, as I already had been working at a military lab since I was in HS, I asked every advisor I interviewed with if research work I was doing at the lab could be the basis of the phd if I wanted (i.e. to do the same thing as the labmate mentioned) and almost everyone I spoke to had zero issues with it, with the proviso that the work was appropriate for a phd.

At the end of the day, I did it all on campus (with many summers spent at labs or industry), but a good chunk of the work I did could easily have been done with proper guidance in an industry setting as well (though that's more in retrospect, I wouldn't have known how to sell the ideas to industry back then).