Worked on enhancing real-time log streaming for Ray workers within Jupyter notebooks by addressing delayed or missing logs caused by ipykernel’s stdout and stderr buffering. In the pinterest/ray repository, implemented a solution in Python that ensures immediate log emission by adding flush=True to print statements in the worker log path, aligning notebook behavior with terminal output. The update improved observability and reliability for Ray tasks executed in Jupyter environments, making it easier to diagnose issues promptly. Manual testing was performed to validate the fix, and code quality was maintained through linting scripts and signed-off commits, demonstrating attention to robust backend development practices.