For those of us who are working on serverless functions, we recenly did a writeup about how to use Neo4j drivers with Functions as a Service (FaaS) -- in particular google cloud functions, but the same concepts apply to AWS Lambda.
I've used this technique a couple of times to capture data coming in from webhooks and put the data into neo4j. Combining this with APOC and node refactoring techniques, it's a basic method you can use to set up streaming graph architectures. Messages coming in on a queue (such as kafka, rabbitMQ and others) can be handled by a serverless function, which puts the data into Neo4j. It's the basis of a streaming data import solution.