Alexandre Caulier enhanced the Speech-to-Text Finetune module in the NVIDIA/NeMo repository by introducing explicit type annotations throughout the codebase. Using Python and leveraging type hinting techniques, Alexandre focused on improving code readability and enabling more robust static analysis. This work targeted the speech_to_text_finetune.py file, ensuring that function parameters and return types were clearly defined, which facilitates easier onboarding for new contributors and reduces future maintenance overhead. While the contribution was scoped to a single feature over one month, it demonstrated a methodical approach to code quality and maintainability, aligning with best practices in machine learning software development.