Serverless query processing systems like AWS Athena or Google BigQuery reduce management overhead by obviating the need to provision resources but can become prohibitively expensive without careful management of the total cost of operations. To address the need for better resource management; Sen, Roy, and Jindal propose an end-to-end framework that predictively allocates the optimal computational resources given price and performance objectives. They integrate this parametric modeling framework into AutoExecutor, a system that selects the near-optimal number of executors and cores for SparkSQL queries that run on Azure Synapse.