Semantic Platform Engineer

Role : Semantic Platform Engineer

Duration : Long Term Contact

Location: Boston, MA (Onsite)

• Client is building Energy Intelligence for Services: an AI-driven service that turns existing condition-based maintenance (CBM) analytics into normalized health scores, contextual diagnostics and prioritized action plans for the critical power equipment of data centers, starting with UPS and busway.
• The service is developed and validated on historical, already-diagnosed customer data rather than on live connections. A semantic layer - an asset and relationship model built on Company's existing ontology work - gives the diagnostic agents the context they need: which assets are connected, what depends on what, and where they sit on site.

Key Responsibilities:

• Deploy the reused Client ontology engine and a single semantic repository shared by the partner teams.
• Build the pipelines that load the resolved data into the repository on Azure, alongside the existing CBM analytics in Databricks.
• Keep the deployment portable (containers, infrastructure as code) so it can later move to another cloud without redesign.
• Expose the graph to the diagnostic agents through documented, read-only query endpoints, and measure query performance.
• Treat the ontology and mappings as code: versioned, tested (competency questions as SPARQL tests, SHACL conformance) and run in CI.
• Write the documentation, schema summaries and example queries that ground the diagnostic agents.
• Maintain the competency questions and expected answers used to test the agents on historical incidents.

Mandatory Skills:

• 4+ years building knowledge graphs, ontologies or semantic data platforms that were used in production.
• Hands-on RDF/RDFS, OWL 2, SPARQL 1.1 and SHACL.
• Solid Python (rdflib, pySHACL or similar), git and CI.
• Ability to work with domain experts and to explain modelling choices to software engineers.
• Deploying and operating an RDF triple store or reasoner (for example GraphDB, Stardog, RDFox, Apache Jena/Fuseki).
• Azure: containers, infrastructure as code, CI/CD; experience working with Databricks.

Preferred Requirements:

• Electrical power-chain or data-center domain knowledge (UPS, busway, switchgear, PDUs) and maintenance vocabulary (failure modes, FMEA, condition monitoring).
• Standard vocabularies: Brick, RealEstateCore / ASHRAE 223, SOSA/SSN, QUDT, ISO 14224, IEC 81346.
• Combining knowledge graphs with LLMs: GraphRAG, text-to-SPARQL, ontology-grounded prompting.
• Property graphs (Neo4j/Cypher), with an informed view of when each approach fits.
• Protégé, TopBraid or LinkML used at scale.