[D] Seeking Advice - For graph ML, Neo4j or nah?

Believe my concerns are fairly general so would appreciate general opinions as well as expert advice, if such is forthcoming.

I'm working on a project to implement a knowledge graph, and the important requirements are:

1. Every node needs an embedding
2. The graph needs to be persistent, because people are adding things to it fairly regularly.
3. The graph is going to ingest data constantly
4. The graph needs to be updating embeddings, inferring connections and missing properties, pretty …

advice graph machinelearning ml neo4j