Better way to integrate Kafka with Akka Cluster Sharding

Question

We have Kafka as the bus and Akka Cluster Sharding as the application distributed cluster. So we need to consume data from Kafka and process them in Akka Cluster.

For a now we implement separate module which is Kafka consumer and Akka Cluster client node. It reads message from Kafka and with self-written guarantee delivery functionality send it to Akka Cluster waiting for self-written technical acknowledgement.

It works more less good. But it seems there should be something that provide such mechanism out of the box or already written tested library. I took a look at Alpakka but couldn't clearly understand does it provide such functionality. Does anyone have good experience with it? Is it what we need for such types of communications?

Does anyone have the same necessity? How do you resolve it?