As far as a web site is concerned: you can Google as well as me -- and a cursory search doesn't seem to turn up a site such as you described. Simply because of the scope and ambition of your project, it is not especially surprising that there is not a large community dedicated to doing it.
Perhaps a good way to create a demo portfolio would be to start such a website?
I should caution you that TAOCP is not the end-all and be-all of algorithms, datastructures, and computer science in general -- the field is too broad for one person to know it all. You don't need to be "versed in TAOCP" -- it sounds like you want to learn algorithms and datastructures, and ideally, to be able to find a job doing them. (I will also warn that finding a job where you can spend a significant amount of time doing algorithms is much harder than it sounds, even if you are both qualified and certified...)
That said, your dedication is very impressive. I have read through TAOCP, but I have never even considered working through it from beginning to end, creating demos, etc. Such a project would take years, even if you already knew all the material! Though, at the end, you would certainly know a lot about algorithms, and learn quite a lot of math, as well.
So, in preparation, if you haven't done much in the way of algorithms before, I would recommend Introduction to Algorithms, by Cormen, Leiserson, Rivest, and Stein. It's a good, modern text on the subject, with clear explanations and lots of substance. It may be thick, but it is nowhere near as dense as TAOCP. There are plenty of other texts available, some of them freely available on the web, but Introduction to Algorithms is popular, and for good reason.
In case you haven't tried it yet, I would also recommend skimming through TAOCP without worrying about understanding everything. Just try to get a feel for it -- go ahead and even skip over sections that are too esoteric, as long as you have some better idea what they're talking about than you had before. Note that, even if some sections are opaque, other sections are relatively approachable.
Also, visit a good research library and try following some of the references -- from TAOCP, or from whatever other texts you are using. It's not actually that hard, and tracking down and using technical papers is a valuable skill.
Finally: awareness of how much you don't know is normal, and learning more will increase this awareness, not diminish it. The world is too big to understand everything in it -- even in the relatively constrained field of algorithms and datastructures. But, by looking around and making yourself aware of what's out there, you take the first steps towards exploring it in depth.