Hello everyone
I am a recently graduated engineer who is currently working as an SDE and I wanted to start contributing to pytorch as a hobby and have what I would consider decent experience with CUDA for someone so early in his career.
I have worked in writing CUDA code for solving graph problems in my college as a semester project and have experience with it in my summer internship in KLA tencor.
As such I have a solid foundation of synchronization primitives, threading and basic performance consideration for GPU/SIMD based program execution.
I would like to start contributing to Pytorch consistently with a time budget of ~10hrs per week and as such would really appreciate mentorship / a guide on how to start understanding the backend GPU codebase properly .
Thanks in advance.