[kornish]

It's also worth noting that with Spark, you can perform arbitrary computation using the Dataset API and operating on case classes.