Worked on expanding multilingual text-to-speech capabilities in the NVIDIA/NeMo repository by delivering production-ready Hindi (hi-IN) support. Developed a Hindi character tokenizer with dedicated grapheme and IPA character sets, updating locale configurations to enable accurate Hindi language processing. Leveraged Python and natural language processing techniques to implement language-specific tokenizer rules, ensuring robust handling of Hindi text. Established comprehensive unit testing to validate the correctness and stability of the new tokenizer, supporting reliable deployment in production environments. This work enabled Hindi text-to-speech workflows, facilitating broader localization and unlocking new opportunities for Hindi-speaking users in the NVIDIA/NeMo ecosystem.