In recent years there have been a lot of interest in machine learning compilers (they are also called deep learning compilers or AI compilers). Yet still they are a niche topic even within the niche field of compiler engineering. This post will try to make things clear about the nature of their function and their design and how they are different from “tradition” compilers. This post will also introduce relevent context and background information needed to build a proper intuition. The reader is assumbed to have some background or familiarity with machine learning and deep learning and compilers but I will try my best to give enough context to grasp it.