To tell the truth, the synthesis of complex compounds is the easy part of drug discovery. You learn a whole lot of different reactions and use them to build up structures like tinker toys. That isn't to say that it doesn't take a lot of time and effort, but it is at least relatively straightforward.
The "knowing what compound to synthesize" is the problem. This is driven a lot more by biology than by chemistry. It all comes down to "what are you trying to accomlish with the compound." And by that, I don't mean something that can be answered by anything as simple as "cure cancer". I mean, what specific biological transformation are you trying to modify, and how much does it need to be modified? The whole "disease pathway research" box is trying to find a single biological process which, if it was sped up or slowed down, would modify a disease. Of all the millions of different biochemical reactions constantly occurring inside an organism, we want to affect just one, in a very precise manner, and sometimes even only in a very specific part of the body. It's like trying to throw a handful of sand into a clock, but getting just exactly the right shape and size of sand grains that it will only affect one gear of the clock and will slow that gear down just enough that the clock will keep perfect time instead of running half a minute fast.