How can I prepare for a PyTorch contributor position at Meta?

To be clear, are these full-time roles on the PyTorch team at Meta? Since that is very different from an invitation to contribute to an open-source repository. The requirements and barriers to entry will, of course, be very different.

In either case, though, I like your idea to contribute to open source, and perhaps even on PyTorch directly.

One of the insights I learned from James Pearce in his course "Become An Open Source Master" is that you don't need to contribute to the biggest, fanciest project. There will likely be many other contributors, so it'll be hard for you to earn the trust and scope you want to show off your skills. You want to get a lot of "reps", meaning that you should be able to make many PRs that get accepted and merged in.

You will likely have an easier time doing that by contributing in a smaller, less popular project in the same ecosystem. I'm sure there are many projects connected to PyTorch in some way. I highly recommend How To Evaluate A Project’s Health.

Aside from that, do all the normal ML Engineer interview prep, such as Ilya's course about ML System Design.