AIP is not an execution platform, so it does have any parallelism explicitly built in. Instead, it can take advantage of whatever parallelism is offered by the JVM environment.
A simple mechanism for enabling parallel execution is to implement Producer classes whose output type is a Future[_]. Downstream Producers that consume Future results can use Await to block on inputs if needed or use Future.map() to queue execution without blocking.
For distributed execution, support for running pipelines using Apache Spark can be found in the spark sub-project.