I don't think it would be possible to robustly convert an R script using base R to the equivalent {data.table} commands.
If you want your R developers to have the speed of {data.table} without learning its syntax, you could try something like {tidytable}, which uses tidyverse syntax that is converted to {data.table} commands.
That being said, I'd encourage anyone that regularly has to analyze large data sets to invest in learning {data.table}. As a long time {data.table} user, I think it is very worth the time investment.