Read what I said again, the meaning of it is that you're mixing time-periods, a better example of this is like comparing the modern English speakers to an the Anglo-Saxson speakers from 1000 years ago, while modern English does indeed come from Anglo-Saxon, the two languages are NOT the same, this is what Old English sounded like:
http://www.youtube.com/watch?v=mVyXDYp60BE
Do you understand? No.
Proto-Indo-Iranian is also a different time-period, the Proto-Indo-Iranian is connected with the Andronovo culture, not Mitanni.
Sanskrit clearly comes from Indo-Aryan, Mitanni (Based on whatever words were found), also shares this Indo-Aryan connection, which I should add, was mostly just the names of Gods that were worshiped by the Vedic Aryans (Sanskrit speakers), so in reality, there's no strong evidence on whether the Mitanni spoke an Indo-Iranian at all, but if they did, it would have been comparable to the Vedic Aryans that migrated to India, not the language of the Medes/Persians.
I accept the Kurds as an Iranian West Asian people, meaning the majority of their genes are native to West Asia but they carry some Indo-Iranian genes, if anything, I think the ancient Hurrians and other similar ancient West Asian populations probably played a bigger role in Kurdish DNA than the ancient Iranians, this has been proven by the autosomal DNA which shows mostly West Asian.
I don't try to link the Kurds with anything, but if you were to ask, the Kurds are closer to some of their non-Indo-European neighboring populations such as Turks, Assyrians, Lebanese etc than they are to actual Indo-Europeans such as the Russians or Ukrainians for example, this goes to show you that just because certain groups speak a certain language, it does not mean much in genetics.