AI Expert-Knowledge Graph

Key Responsibilities
Knowledge Graph Architecture & Design

• Design and implement enterprise-grade knowledge graphs on Microsoft Fabric, leveraging OneLake, Direct Lake mode, and semantic model layers.
• Develop ontologies, taxonomies, and entity-relationship models aligned to business domains (financial services, healthcare, aviation).
• Architect graph-based data products integrating structured (SQL, lakehouse), semi-structured (JSON-LD, RDF), and unstructured sources.
• Define standards for graph schema governance, versioning, and lineage tracking.

• Build knowledge graph pipelines using Microsoft Fabric Data Engineering (Spark, notebooks, lakehouses) and Data Factory.
• Integrate Azure AI Search knowledge store with graph indexes for semantic retrieval and GraphRAG patterns.
• Connect graph models to Power BI semantic models via Direct Lake for real-time, governed analytics.
• Leverage Azure OpenAI and Fabric Copilot capabilities for natural language querying over knowledge graphs.
• Implement Fabric Real-Time Intelligence (Eventstream, KQL) for dynamic knowledge graph updates.

• Design GraphRAG (Graph Retrieval-Augmented Generation) pipelines that ground LLMs with structured enterprise knowledge.
• Build agentic AI workflows using graph traversal for context-aware reasoning (e.g., Azure AI Foundry, AutoGen, Semantic Kernel, MCP-enabled agents).
• Develop entity resolution and disambiguation pipelines using NLP and graph embeddings.
• Design and expose knowledge graph capabilities as MCP (Model Context Protocol) servers, enabling LLM agents to query, traverse, and update graph entities through standardized tool interfaces.
• Implement graph-based fraud detection, risk scoring, and compliance reasoning patterns for financial clients.

• Align knowledge graph assets with enterprise data governance frameworks (Informatica EDC/Axon, Microsoft Purview).
• Model regulatory and policy knowledge graphs for GCC compliance frameworks (PDPPL, QCB, CBUAE, NPC).
• Publish and steward graph-based business glossaries and data lineage maps integrated with data catalogs.
• Enforce data quality rules and semantic constraints within graph structures.

• Lead client workshops to elicit domain knowledge and translate it into graph ontologies and use cases.
• Produce architecture blueprints, technical proposals, and proof-of-concept demonstrations for pre-sales engagements.
• Mentor junior data engineers and solution architects on graph modeling best practices.
• Collaborate with data scientists, AI engineers, and enterprise architects on cross-functional AI platforms.

• Cypher (Neo4j), Proficiency in graph query languages:
• SPARQL (RDF/OWL triple stores — Apache Jena, AWS Neptune, Stardog)
• Gremlin (Azure Cosmos DB Graph API, TinkerPop)
• OpenCypher (Memgraph, Amazon Neptune Analytics)
• Hands-on ontology engineering using OWL, RDFS, SHACL, and SKOS standards.
• Experience with property graph and RDF graph modeling paradigms and trade-off analysis.

• Demonstrated experience with Microsoft Fabric end-to-end: OneLake, Lakehouses, Warehouses, Data Pipelines, Notebooks.
• Familiarity with Fabric GraphQL API for exposing knowledge graph entities as managed data products.
• Building semantic models in Fabric for Power BI Direct Lake and AI-ready data assets.
• Integration with Microsoft Purview for unified data governance across Fabric workspaces.
• Experience with Fabric Copilot, AI Skills, and Real-Time Intelligence workloads.

• Azure AI Search — knowledge store, indexers, semantic ranking, vector hybrid search.
• Azure OpenAI Service — prompt engineering, GraphRAG, function calling for graph traversal.
• Azure Databricks or Fabric Spark for large-scale graph processing (GraphX, GraphFrames).
• Azure Data Factory / Fabric Data Factory for ETL pipelines feeding knowledge graphs.
• Azure Cosmos DB (Gremlin API) for operational graph workloads.

• Python (RDFLib, NetworkX, PyKEEN, SPARQLWrapper, LangChain graph tools).
• PySpark for distributed graph data processing on Fabric / Databricks.
• Experience with NLP pipelines for entity extraction and relationship mining (spaCy, Azure Language Service, GLiNER).
• Knowledge of vector databases (Azure AI Search, Qdrant, Weaviate) for embedding-based graph enrichment.
• Familiarity with Model Context Protocol (MCP) for building graph-backed tool servers that expose SPARQL/Gremlin query capabilities, ontology lookups, and entity resolution as callable tools for LLM agents (Claude, GPT-4, Azure AI Foundry).