R is a statistical programming language. Although in general we prefer python, stable versions of some statistical models are implemented only in R. R also has great tools for working with tabular data.
Although R is a great language -- and in fact, it's the best overall tool for a lot of use cases -- there's not a lot that you can do in R but not Python or Julia. That means that developing infrastructure in R is generally not a good use of our time.
However, R is widely used in some communities -- including corners of the stats community -- so you may end up using it for a collaboration. Alternatively, you may find that there's some model you want to run that's already implemented smoothly in R. In that case, don't reinvent the wheel!