NASA doesn't build much. The only reason their employees do any engineering at all is so they stay somewhat competent enough to write requirements and evaluate contracts. Most of the money and work has been, is done, and will continue to be done by contractors.
Take Apollo. North American made the command and service module and second state, Grumman made the LM. The Saturn V first stage was built by Boeing with Rockedyne Engines. The third stage was built by Douglass, The avionics by IBM.The escape system by Lockheed.
Now since most of the integration work can be done by industry NASA has started to even back out of the integration role. Let the contractors build the whole thing. This doesn't mean NASA doesn't do anything. SpaceX, Orbital, etc uses NASA technology and experts all of the time. The difference is these companies go to NASA to ask for help where in the past the contractor had to pass NASA design reviews which slowed things down and made everything cost more.
I for one like this direciton. Launching rockets is proven technology. It's time for business to figure out how to make it economical. With these savings NASA can spend more on more payloads. If launch costs go down significantly then the spacecraft costs will drop as well. You don't launch a $10M spacecraft when you have to pay $200M for the rocket. But if the rocket costs $10M then you just might. 20 times the missions even if a bunch fail is still quite a bit of science.