Case study · Data ConnecXion
Graph analytics platform (Neo4j)
Relationship-shaped questions relational databases answer slowly or never.
Problem
Relationship-shaped questions relational databases answer slowly or never.
Approach
Helped design and build a graph-based data analytics platform on AWS: DynamoDB and PySpark for data processing, Neo4j for graph storage and analytics, serverless stack provisioned with Chalice. Deployed and managed Neo4j cluster infrastructure, developed its APIs, architected a long-jobs executor microservice with RedisRQ, and implemented graph-ML algorithms using Neo4j Graph Data Science and PySpark GraphFrames. Certified Neo4j Professional - this is where that certification did real work.
Architecture
When relationships ARE the data
Some questions are joins; some are paths. Fraud rings, ownership networks, dependency chains - relational databases answer these slowly or not at all. Graph-ML over Neo4j turned relationship-shaped questions into first-class analytics, and the serverless AWS stack kept the platform light to run.