AKKA FSM in spark streaming

Question

I have a requirement to create state machine implementation in my spark streaming application. After reading thru some posts found AKKA comes with FSM out of the box. I created a simple AKKA FSM and I am able to run it locally. I am not sure how to integrate this code into spark structured streaming (mapGroupsWithState). I want to take this FSM and plug into mapGroupsWithState so I can store the last state even if my application goes down.

object BuilderFSM {

//States
sealed trait MachineState
case object tuned extends MachineState
case object trick extends MachineState
case object nonTuned extends MachineState

case class TuningEvent(macId: String, acctNumber: String, eventType: String)

}

object InputEvent {
sealed trait tuneEvents
sealed trait trickEvents

case object PLAY extends tuneEvents
case object REPLAY extends trickEvents
case object PLAY1 extends trickEvents
}

class BuilderFSM extends FSM[MachineState, TuningEvent] {
startWith(nonTuned, TuningEvent("DEVICE", "ACCOUNT", "NOTHING"))

when(nonTuned) {
case Event(PLAY, _) => {
  println("I am in NON-TUNE -> Going to Tuned")
  goto(tuned)
}
case Event(REPLAY, _) => {
  println("I am in NON-TUNED -> Going to Tuned")
  goto(tuned)
}} initialize()
}