Major background noise reduction

Unfortunately, as you have discovered, NR only works on continuous background noise, and even using the Sound Model tool will leave you in the same situation with the number of times you have to use it - and that doesn't work too well with background sound either. If your interview was recorded in stereo with the voice central and everything else off to one side, then the Center Channel Extractor might help - you could boost the center signal and effectively reduce the sides.

Basically though, you can't realistically do anything effectively in this situation. Often you are better off leaving it as it is, as anything you remove also removes part of the wanted speech signal because it's all in the same frequency band. Even though it's noisy, it's still usually easier to undstand like this. Please don't fall into the trap of believing what you saw on CSI - their extraction stunts weren't real. There are a couple of tricks you can do with iZotope's RX, but even those are more aimed at a forensic extraction, and won't sound good - certainly not what you'd want for an interview. Bottom line - I'm afraid no software is going to fix this for you - sorry.