Avanade interview question

How would you optimize a large dataset transformation pipeline in Databricks for performance? Python Leet code SQL query