This is really neat. We have some Kafka transformation/deduplication use-cases that we can leverage this on.

Thanks for sharing!