A team of data engineers has developed a Databricks Job in production with multiple tasks. Their goal is to minimize costs related to compute resources and optimize the execution time for each task. What is the best strategy to achieve these objectives?