Worked on the aws-mwaa/upstream-to-airflow repository to deliver EMR Container Cancellation on Task Kill, enhancing Airflow’s integration with AWS EMR. Developed a cancel_on_kill flag within the EmrContainerOperator and EmrContainerTrigger, enabling automatic cancellation of EMR container jobs when tasks are killed in deferrable mode. Updated the EMR job execution logic in Python to ensure timely resource cleanup during deferral and timeout scenarios, reducing wasted compute and improving workflow reliability. Added comprehensive tests to validate the new behavior, increasing confidence in production deployments. Leveraged AWS, Airflow, and back end development skills to improve resource efficiency and job management control.