The cofounder of Symbolics, Dan Weinreb, apparently works there as well. They also seem to do a lot of Lisp programming. Seems like a pretty cool company.

  >> seem to do a lot of Lisp programming

However, the puzzles seem weighted to requiring a solution using Java.

And according to Dan Weinreb's blog, they will most likely not use Common Lisp again on another ITA project:

http://danweinreb.org/blog/rumors-of-ita-acquisition-are-jus...

That's a shame.

I spent quite some time working on the "palindromic pangram" problem. I believe I solved it, but my program took several days to finish, so there's probably an optimization I'm still missing. If my program is correct, the shortest palindromic pangram is length 57. Has anyone here done better (either a shorter solution or faster running time)?

How can one improve in these sorts of problems? Crack open CLRS and read that?

I see they're taking a hint from Google.

What do you want to imply? I do not see any connection to Google.

ITA had those puzzles for quite a long time. (Perhaps before Google was even founded?)