We are building the core memory and retrieval layer for agentic AI systems operating over large-scale enterprise data. This role focuses on designing and operating high-quality data pipelines that move data from a cloud lakehouse into a Neo4j-backed semantic graph optimized for agent reasoning and GraphRAG retrieval.

This is a hands-on data engineering role, prioritizing pipeline automation, graph construction, performance tuning, and reliability over academic ontology design. Your work enables agents to retrieve accurately, reason reliably, and operate with enterprise-grade context and traceability.

Key Responsibilities

• Build and maintain ETL/ELT pipelines from Microsoft Fabric (or similar) into Neo4j
• Transform structured and unstructured data into clean graph models (nodes, edges, metadata)
• Implement automated ingestion, delta updates, and streaming/CDC patterns
• Implement and maintain Neo4j graph structures, indexes, constraints, and performance tuning
• Develop Cypher queries, APOC procedures, and ingestion utilities
• Enable GraphRAG structures (entities, chunks, embeddings) for high-quality retrieval
• Optimize graph-based retrieval for agent workflows (hybrid search, entity linking)
• Ensure data quality, lineage, auditability, and monitoring
• Collaborate closely with AI Architects, Agent Developers, and UX/AX teams

Required Skills & Experience

• Strong Data Engineering background (pipelines, orchestration, modeling)
• Hands-on Neo4j experience:
  • Cypher, APOC, graph modeling
  • Bulk ingestion, indexing, performance tuning
• Experience with Microsoft Fabric, Synapse, ADF, or similar cloud data platforms
• Familiarity with GraphRAG, retrieval systems, or RAG hybrids
• Comfortable with Python or TypeScript for pipelines and APIs
• Understanding of LLM behavior and agentic retrieval workflows (preferred)

Nice to Have

• Experience in regulated enterprise domains (banking, finance, operations)
• Familiarity with vector databases and embedding pipelines
• Experience supporting multi-agent or AI-driven systems
• Exposure to semantic models, ontologies, or knowledge engineering

‍