Question 1

What does Neo4j consulting include?

Accepted Answer

Graph model design (node-vs-property modelling, relationship cardinality, supernode mitigation), Cypher query strategy (traversal patterns, index design, query plan tuning), Graph Data Science (GDS) algorithm selection for fraud / recommendation / community detection, AuraDB tier sizing vs self-managed economics, knowledge-graph + vector hybrid retrieval architectures, and RDBMS-to-graph migration planning. Deliverables are written recommendations and topology proposals.

Question 2

How do you design a graph model for production?

Accepted Answer

Three modelling decisions drive most graph performance: (1) what becomes a node vs a property — high-cardinality attributes that participate in traversal become nodes, scalar metadata stays as properties; (2) relationship direction and cardinality — Neo4j's traversal cost depends on these decisions; (3) supernode mitigation — nodes with millions of incoming edges (a popular product, a major hub-airport) need property-based filtering or relationship sub-typing to avoid pathological queries. We audit the proposed model against expected query patterns before code ships.

Question 3

When does GDS pay off over running algorithms externally?

Accepted Answer

GDS pays off when: (a) the algorithm runs on the same graph repeatedly with different parameters, (b) the graph fits in heap memory of the GDS-runtime nodes, (c) you need the algorithm output written back to the graph for downstream queries. Run algorithms externally (Python/Spark/NetworkX) when: the graph is small enough that export/import is trivial, the algorithm runs once for analysis, or you need a Python-ecosystem ML pipeline. We model the workload before recommending.

Question 4

AuraDB tier sizing — how to pick?

Accepted Answer

AuraDB Free is for dev/learning (no production). AuraDB Professional starts at 1GB and scales — fits most production workloads with steady throughput. AuraDB Business Critical adds multi-zone HA, point-in-time recovery, and minimum SLA. AuraDB Virtual Dedicated is single-tenant for compliance / data-residency requirements. We size against your graph size, query throughput, and recovery requirements — and compare against self-managed Neo4j Enterprise on K8s for sustained workloads.

Question 5

Migrating from RDBMS to Neo4j — what's realistic?

Accepted Answer

Schema-to-graph mapping is the hardest piece — relational tables become nodes, foreign keys become relationships, and properties move from columns to node/edge attributes. The hand-built mapping rarely survives review; we usually iterate against expected query patterns. Data movement via the LOAD CSV or APOC tooling. Application tier needs rewrites to use Cypher and the Neo4j driver. Most mid-size RDBMS → Neo4j migrations land in the 8-16 week range — schema design dominates the timeline.

Question 6

Knowledge graph + vector — how do they combine?

Accepted Answer

RAG architectures benefit from hybrid retrieval: vector search for semantic similarity, graph traversal for relationship-aware context. Neo4j 5.x added native vector indexes (HNSW), so you can store embeddings alongside graph nodes and combine vector kNN with Cypher traversal in one query. This is significantly more powerful than pure vector retrieval for domain-aware RAG. We design the hybrid pipeline — embedding model, vector index, traversal pattern, and the orchestration that ties them together.

Question 7

Does consulting include migration execution?

Accepted Answer

No — execution lives on dedicated engagements. Consulting is the decision phase: design the graph model, plan the Cypher strategy, pick the GDS algorithms, size the deployment, decide AuraDB vs self-managed. Execution covers RDBMS → Neo4j cutover, GDS pipeline buildout, AuraDB onboarding, or operational handoff. The split keeps consulting deliverables crisp — written recommendations, not code-shipping work.

Neo4j Consulting Services

What our Neo4j consulting covers

Graph Model Design

Cypher Query Strategy

GDS Algorithm Selection

AuraDB vs Self-Managed

Fraud & Identity Graphs

Knowledge Graph + Vector

RDBMS → Graph Migration

How a Neo4j consulting engagement is shaped

Graph Model Review

Migration Strategy

GDS Pipeline Design

Greenfield Design

Neo4j consulting — common questions

Ready to make the call on Neo4j?