I'm not familiar with the synthesis development of remdesivir, but typically those side groups are discovered through a process called Structure Activity Relationship (SAR). The basic idea is you start with a bunch of different compounds (like 1000's) and look at their ability to interfere with a specific enzyme in a high-throughput screen. When you rank the 1000's of compounds, some will work better than others. Then you compare the structures of the top hits and look for similarities. So in this case those huge groups attached to the adenosine likley improve binding to the enzyme because they're greasy/hydrophibic.

So then you add different greasy groups to new compounds and screen those. So will be worse, but sometimes some will be better. Then you look at the better ones, like having a nitrile group off the 1' position of the ribose and maybe that started as an amine (I'm making shit up here) and they decided to make it stick further out (IDK).

Anyway, I did some quick looking at it seems like remdesivir is a prodrug that gets modified by other enzymes to become triphosphorylated and then incorporated into the RNA genome of the virus (https://www.nature.com/articles/nature17180/figures/1). So they got super lucky finding it! Check out that paper for the story.