My project (the Pacific International Training Desk) is a cooperative agreement between the US National Weather Service and the University of Hawai'i. We have an interesting partnership beyond that, here in Honolulu, because the NWS Honolulu Forecast Office is located on the campus of the University of Hawai'i, in the same building as the Atmospheric Sciences Department. That is a relatively unique arrangement, but it leads to a lot of interaction between professors, students, forecasters, and my training participants from Pacific-Island met services. Students and faculty attend twice-weekly map discussions by the forecasters, university seminars include the met service, there are student internship opportunities in the forecast office, and many of the forecasters earned a degree or two from the university. There's nothing like co-location to foster a lot of interaction - formal and informal - between the met service and a university, and I realize that isn't practical everywhere.
We also have a scientific officer who leads research-to-operations transitions. I think researchers benefit from spending time in operations to get a better understanding of how tools and knowledge are used in time-critical decision-making situations, and how new tools and knowledge can be integrated into existing systems and workflows. Having a focal point to coordinate that is helpful.