Contributed to the livekit/agents repository by developing analytics-oriented features and addressing targeted bugs in the speech and TTS pipeline. Built audio duration reporting for speech recognition sessions, enabling enhanced usage tracking and deeper insights into session lengths and user engagement. Utilized Python and asynchronous programming to implement robust backend solutions, focusing on API integration and reliable code paths. Addressed privacy concerns by preventing leakage of function call text into TTS output, setting content to None, and expanding unit test coverage to ensure sensitive information remained protected. The work emphasized business value through improved analytics visibility and strengthened content privacy safeguards.