Thankfully I didn't need very deep EE/radio knowledge to get up to speed in my role. It's sufficient to be able to read an EE paper and extract out the end result. For example: one of our hardware guys wants me to simulate a phased-array antenna. They hand me a paper that derives the gain calculation based on the geometry of the array. I don't need to understand the derivations as long as I can translate the result into modelling software.

On the software/networking side of things the biggest change is working with protocols other than http.