so inputKafka --> DataJob (scala pipeline) --> outputKafka and DataJob: inputType --> middleType --> OutputType (all these are scala types which are represented as Datasets)