Merge smaller batch jobs into bigger jobs

[nbgl]

Similar to #17. Suffers from the same issues.

A user may choose to upload data in tiny batches (say, one row per request). How do we make sure that we don’t end up with suboptimal batch sizes when we’re performing the hashing?

Aha! Link: https://csiro.aha.io/features/ANONLINK-88