Worked on the Lightning-AI/litData repository to enhance the streaming module by implementing dynamic downloader registration, enabling future plug-in downloader extensions. The approach involved introducing a register_downloader function and refactoring the API for clarity and consistency, specifically renaming get_downloader_cls to get_downloader. Comprehensive unit tests were added to validate the registration flow and ensure the accuracy of AI-generated docstrings, supporting maintainability and code quality. The work was carried out using Python and focused on API design, software engineering best practices, and testing. This contribution improved the extensibility and maintainability of the streaming pipeline without addressing bug fixes during the period.