Can anyone comment on how Ray fits into the current distributed systems landscape? How does it compare to Spark, Hadoop, Akka, distributed Tensorflow, etc.? It

Can anyone comment on how Ray fits into the current distributed systems landscape? How does it compare to Spark, Hadoop, Akka, distributed Tensorflow, etc.? It would be really helpful to have an FAQ on the github.io page that answers the question "When should I use Ray vs. X".