Zvika improved the reliability of dataset file processing tests for the TonicAI/textual repository by developing and integrating a shared wait utility into the test suite. Using Python and pytest, Zvika refactored existing test utilities to centralize timing and synchronization logic, ensuring that background file processing completed before tests proceeded. This approach addressed race conditions that previously caused flaky test failures, resulting in a more stable continuous integration pipeline. The work focused on backend development and test engineering, enhancing maintainability and accelerating feedback for dataset-edit workflows. Zvika’s contributions deepened the quality and robustness of automated testing for dataset-processing features.