Save dataset attachments for data engineers before pipeline runs

Your inbox dataset attachments arrive in threads, delaying model training and ETL jobs. They are saved into a shared processing folder so data engineers can trigger the next scheduled run.