[noxs]

this is quite the difference I saw among different ML Engineers during my work.. most can only do python (more or less like a frameworker) and some amazing ones can go all the way down to the CUDA to squeeze out the performance.