I recently gave a webinar about contributing to ArviZ, that is now available on YouTube which I believe can be of interest.
ArviZ is a Python package for exploratory analysis of Bayesian models. It integrates with pyro and numpyro and can be used, among others, for tasks such as visualization, diagnostics or model comparison of their models.
In this webinar we go over both social and technical aspects we face when we contribute to ArviZ and to open source in general. We cover: finding an issue to work on, understanding how to work on it, to submitting the pull request and addressing the feedback received, and challenges faced. The talk will be focused on ArviZ, but it should also be useful to anyone interested in contributing to open source.
Moreover, in case someone is interested in both ArviZ and Pyro we have some open issues that could make the integration even better and might be a good starting point: