Let's not forget the deltaVee penalty because the thing has virtually no gravity well to skate onto. You have to do it the hard way, firing rockets to match Ceres' solar orbital velocity. Works out to be 9kps to get there, as opposed to about 5kps for Mars orbit (excluding aerobraking). This is a nuclear mission, no question.
How did you get the 9kps?
Assuming a propellant depot at EML2 like the one proposed by ULA, I get about 1800m/s for trans ceres injection using an impulsive maneuver at earth flyby, and <4.5km/s for matching ceres orbital velocity using an impulsive maneuver during ceres flyby (it still helps a bit).
So a very conservative estimate for chemical propulsion delta-v including landing would be 7km/s.
For the trip back, you would have to manufacture your propellant on ceres. But that should not be too hard given that there should be plenty of water there. Of course you would need either a nuclear reactor or concentrating solar arrays at that distance to the earth. On the plus side, propellant boiloff would not be a big problem even with LH2.
Edit: I still think that phobos or deimos is the best first beyond LEO goal.
Edit again: I still have to find the error in my back of the envelope calculation, but using this
excellent tool (link) from vanilla, it seems that the 9km/s is correct. Too bad.