Describe a time you used PySpark to transform a massive dataset. How did you optimize the job for performance?
