GraphRAG based system

I am looking for a developer with hands-on experience in GraphRAG, Neo4j, Cypher, and Python to help complete and refine an existing project related to Knowledge Graphs.

The project is already implemented and approximately 60–70% complete. The Knowledge Graph has been built in Neo4j, data has been integrated, and most backend functionality is already available. This is not a project from scratch. The work mainly involves debugging, refinement, optimisation, and improving retrieval quality.

Key tasks include:

• Reviewing and cleaning existing Cypher queries
• Fixing schema-query inconsistencies
• Improving the Query Classifier and entity detection
• Enhancing GraphRAG retrieval and response relevance
• Debugging existing Python backend code
• Improving result ranking and evidence retrieval from the Knowledge Graph
• Performing minor Knowledge Graph cleanup where required

The ideal candidate must have prior experience working with GraphRAG systems and should understand how graph-based retrieval pipelines interact with LLMs. Experience with Neo4j, Cypher query optimisation, and GraphRAG architectures is highly preferred.

The codebase, Knowledge Graph, and major functionality are already available. Most of the work involves understanding the existing implementation, identifying issues, and refining the system rather than developing new components from the ground up.

Deliverables include:

• Better GraphRAG retrieval quality
• Bug fixes and schema alignment corrections
• Documentation of major changes

Budget: PKR 8,000–10,000  (negotiable)

Please apply only if you have previously worked on GraphRAG-based projects and are comfortable working with an existing Neo4j + Python codebase.