In com.github.msemys.esjc.util.UUIDConverter.toBytes() UUID is converter to byte array directly, while .NET Guid does some inversion.
See:
As a consequence, if eventId is hard-coded to some UUID (read from string) and such event is saved to event store, EventStore will report a different EventId (string) value (consistent with different byte ordering).
As an example, Java UUID 92651dfa-0d83-53e2-a6f7-e6afad26121e will be shown as fa1d6592-830d-e253-a6f7-e6afad26121e in EventStore (when EventId is checked via EventStore HTTP API).
This is mostly problematic in testing/debugging while real-life use is ok as long as UUIDConverter methods for converting UUID to/from byte array are simmetric.