GSOC 2021 Replace backends with xarray

Hi there,

I am curious about the project listed as replacing backend with xarray listed as one of the gsoc 2021 projects. I am reasonably familiar with how to use pymc3 and arviz and understand the need to work to enable the two to work better together through the use of xarray. However the details of the project specification are fairly light on the github page. I was wondering what level of experience would be required for someone to make a meaningful contribution to this project? Is this project suitable for someone willing to learn more about the backend of pymc3 as they go or does it require a lot of previous experience with working on the backend?

Thanks in advance,

1 Like

Tagging @RavinKumar and @colcarroll


I would say some experience with general python, plus an interest in learning, or a prior understanding, of Bayesian sampling and how that data “looks” is what you need for this project. From what you detailed above it sounds like youd be able to make a meaningful contribution!