How to write CUDA c++ modules for python?

I think my question might be hard to understand, so I will try to explain it here. I wish to delve into the research side of thing in deep learning as was wanted to learn writing in CUDA because I heard that it will be really helpful.


Let's say I wanted to write a matmul library for python. How do I like, build on top of the c++ code in order use it in python? Like, under the hood, [np.dot](https://np.dot) …

cuda cuda c deep learning deeplearning learn modules python research think writing