[Lambda] Create new worker to handle s3 chunks

[AaronVasquez]

Because the worker will need to download from s3, we need a separate worker with different logic.

Future discussion
Should we unify the logic between s3 and badger?