As long as it's R that does the heavy liftting, using one of the many Python frameworks along with rpy2 would make the development easier.</shameless plug>
or you can just do density estimation in python also and not use R at all
There really is a lot of well tested and documented functionality in R that would be boring/time-consuming/inefficient to try to reproduce. I'm not exactly in love with the language, R syntax is fiddly, but I am in love with the libraries.