It (obviously) has all the 8 original Indo-European cases, three numbers (singular, dual, plural) and interestingly three voices (active, passive and medium; however the differences between the active and medium voice wore out quickly, so they both have in fact an active meaning pretty much all the time (just some verbs use one and some the other)), but that's still pretty normal.
However, it has a ton of curious rules (called sandhi) to join adjacent words into one unit, with various changes at the boundaries of the words. For instance, nara(h) tatra ("[a] man [is] there") = narastatra (without spaces and pauses in speech). So in this way, a whole sentence can be written as just one "word", with a bunch of weird changes in the middle. (Also the fact that it's written using a script that has 1 character per syllable does not help, because the words generally join one another "in the middle of a character".) So it sometimes happens that you split the mess in a wrong way, then spend half an hour trying to translate the thing, and then your teacher shows you another splitting that leads to an immediate and simple translation.
Second, it has some crazy composites. It's like German, only several times worse, especially when it comes to composites that just mean several things together. A typical example (with a bad transliteration, so that you can more easily read it ): mrgavyaghrasimharksagajapurne vane = in a forest (vane) full (purne) of gazelles (mrga), tigers (vyaghra), lions (simha), bears (rksa) and elephants (gaja).
And third, it's probably the only language that I know of which has a lot of nouns with several roots. Many words have three of them, like the well-known word "raja": there is the strong root (rāja-), the middle root (rājā-) and the weak root (rajňā-). Each of these roots is used only in certain cases (with number and gender included), so for instance, in singular, masculine and feminine nouns use the strong root for nominative and accusative, but the neuter nouns use the middle root for them; and all of them use the weak root for the other cases. (And it gets more complicated with the other numbers. )
And to make it even worse, it has an awful lot of words. In English, one word has pretty much always more than one meaning, so you don't need to know a whole lot of words. In Czech, the relationship is pretty much 1:1. But in Sanskrit, it's more like 5 words for any single meaning. When I was learning, I knew no less than 6 words for a "king", for instance, all of them being used interchangeably. (And the situation was the same even for some weird words, like "arrow" — I think I knew 3 words for that, now I can recall only two.)
So be glad that you don't know anything about it. But — if you would like some serious mental challenge — you could try to learn it .