A company regularly extracts about 2 TB of data daily from various data sources - including MySQL, MSSQL Server, Oracle, Vertica, and Teradata Vantage. Some of these sources feature undefined or frequently changing data schemas. A data engineer is tasked with implementing a solution that can automatically detect the schema of these data sources and perform data extraction, transformation, and loading to an Amazon S3 bucket. What solution would meet these needs while minimizing operational overhead?