Want to know if this job is worth applying to?
San Francisco, California, United States
Onsite
As an Ontology Engineer on Samba TV's Knowledge Graph & Identity team, you will build, maintain, and extend the knowledge graph schemas, derivation pipelines, and graph data models that underpin Samba's measurement and audience intelligence products. Working closely with the Senior Ontologist and peer data scientists, you will implement ontological frameworks in production, contribute to entity resolution and data enrichment pipelines, and help ensure the graph layer remains accurate, consistent, and production-ready.
This is a hands-on technical role. You are expected to write clean, production-quality Python and SPARQL, take ownership of well-scoped graph work streams, and grow your depth in semantic modeling under the guidance of senior team members.
This role reports to the Data Science Manager, Knowledge Graph & Identity.
Implement and extend Samba's RDF/RDFS/OWL ontology schemas in the graph database - adding entity classes, properties, and constraints in a consistent, governed way under the direction of the Senior Ontologist
Build and maintain SHACL validation shapes for post-load graph consistency checks; identify and triage data quality and schema violations
Support ontology versioning, change log documentation, and consistency checking across schema updates
Write efficient, well-structured SPARQL queries and graph traversals to support downstream data science and product use cases
Contribute to the event-to-ontology transformation and derivation layer - building PySpark/Databricks pipelines that aggregate raw TV viewership and web activity events into durable graph attributes (genre affinity, brand affinity, topic affinity, viewing summaries, lifecycle signals)
Implement derivation logic specified by the Senior Ontologist and data science team; validate outputs against SHACL shapes before graph load
Support incremental refresh and update logic aligned with the graph's batch refresh cadence
Write production-quality Python - clean, well-tested, documented, and reusable by teammates
Work with PySpark and Databricks to process and transform high-volume data as part of graph pipeline development
Apply embedding-based approaches (semantic similarity, vector search) to entity matching and ontology alignment tasks
Contribute to team tooling, documentation, and reusable components that improve knowledge graph development efficiency
Partner closely with data engineering on pipeline design, data quality, and incremental ingestion patterns feeding the materialized graph substrate
Participate in ontology design reviews and cross-functional working groups
Work with product and operations teams to understand use case requirements and translate them into graph schema updates
Actively develop expertise in W3C semantic web standards, RDF-native graph databases, and entity resolution under the guidance of the Senior Ontologist
2–4 years of hands-on experience in knowledge graph development, semantic data modeling, ontology engineering, or a closely related field
Working knowledge of W3C semantic web standards: RDF, RDFS, OWL, and SPARQL - with practical experience querying or building in at least one triplestore or graph database
Familiarity with SHACL or equivalent constraint and validation frameworks for graph data quality
Strong Python skills - clean, readable, production-quality code with testing and documentation
Solid understanding of data modeling fundamentals - entity-relationship design, taxonomies, hierarchies, and how to represent complex real-world relationships in structured form
Familiarity with entity resolution or data matching concepts - understanding of why the same real-world entity appears under different identifiers across data sources
Bachelor's degree required in Computer Science, Information Science, Mathematics, or a related field; Master's preferred
Detail-oriented and proactive about flagging data quality issues and schema inconsistencies
Hands-on experience with Amazon Neptune or Stardog - or equivalent RDF-native triplestore; exposure to data virtualization (Neptune Orion or Stardog Virtual Graphs) a plus
Working knowledge of PySpark and Databricks - particularly for large-scale event aggregation and transformation pipelines
Familiarity with embedding models, vector search, or semantic similarity - applied to entity matching, ontology alignment, or knowledge graph enrichment
Experience with LLM APIs or RAG-based approaches applied to information extraction, entity disambiguation, or schema mapping
Domain knowledge in media, entertainment, or ad tech - content metadata, advertising entities, TV viewership data, or audience/identity data
Exposure to identity resolution, probabilistic record linkage, or device graph approaches