I did a similar thing for our kids (Hindi/English) but we had slightly different goals. Since the kids were getting my English surname, we wanted to give them Hindi first names, but we wanted them to be easily pronounceable in English. So I wrote a character-based ngram language model based on the US census count of names, and used that to score a list of Hindi names. I then improved it by adding a bunch of rewrite rules to collapse similar sounds together to "canonicalize" names before scoring them against the language model.
Finally we ended up with Riya for my daughter (which is pronounced the same as the English name Rhea) and Aarav for my son, which is pronounced the same as two of the most common English words, "are" and "of."
Edit: I was able to find the output from the girl's name list. https://docs.google.com/spreadsheets/d/1Vy1dQunG4iie4H67En9T...