This is pretty much the underpinning question of the entire field of evolutionary developmental biology, so naturally any answer is going to be a bit surface level, and I get out of my depth fairly rapidly to be honest. Still, it is quite interesting.

One of the central ideas is that as an embryo grows, its cells go from being all equivalent multipotent stem cells into being different from each other - at first more specialized types of stem cell that can only turn into certain tissues and gradually specializing more and more. Since these cells are differentiated and expressing different genes from one another, they can then start to co-ordinate with each other using chemical markers and gradients of concentration of those markers across space to regulate what types of cells should be growing/dividing, where in the embryo they should be doing it and at what time they should be doing it.

All of that signaling is controlled by some often complicated networks of regulatory genes - ones which when they are expressed make proteins that selectively attach to other bits of the DNA in that cell and make the genes there more or less likely to be expressed themselves. A lot of evolutionary variation is actually focused on these regulatory systems rather than on the genes which they are switching on and off.

So to my knowledge, something like nose shape likely comes down to some of those regulatory genes controlling where the cells that will eventually be forming the cartilage get placed relative to the skull etc.