I do think finally reading Knuth's work is helping here. As he manages to show just what that means. Far beyond anything I would have ever expected. And, amazingly, it is completely approachable!
Seriously, look at what I did here: http://taeric.github.io/Sudoku.html I still want to clean up parts it. Terribly so. What does this have to do with my day job? Really... nothing. Nothing at all.
Edit: More relevantly, what do my cleanups that I am obsessing over really help with understanding? Probably nothing. I just think I should clean up the function names and such. Stopping to actually understand how the fundamental algorithm works and where else I could/should use it isn't helped by my structuring the code differently.
You may really want to know what code perfection means to you, but you might also want to know what it's like to have a family, or run a marathon, or whatever.
Don't risk your job, but spending the time to make something as good as you can make it is often rewarding.
Not to be all morbid, but the clock is ticking, death comes for us all. If it's bugging you, you should fix it - even if no one else cares. If your own satisfaction isn't enough, you should set it aside and do something worthwhile. (whatever that may mean for you)
Also, I am literally the only person at work that will vote to hire someone regardless of whether they got all of the answers "right."