Senior Ontologist - Knowledge Graph & Identity

**Samba is a media intelligence company. We know what the world is watching, reading, and thinking about — in real time, at scale, across every screen. Our data exists with the consent of over a billion people, organized into the most complete picture of consumer attention ever built. The biggest brands in the world use that picture to make smarter decisions. We think it’s the most interesting data asset on the planet, because it’s the most culturally relevant.** As Senior Ontologist on Samba TV's Knowledge Graph & Identity team, you will own the design, development, and governance of the semantic data models and ontological frameworks that sit at the foundation of Samba's knowledge graph. You are the domain authority for how Samba represents and relates the entities that matter most to our business - and you ensure that representation is rigorous, scalable, and aligned with industry standards. This is a hands-on technical role. You will spend the majority of your time designing ontologies, writing SPARQL, building knowledge graph pipelines, and working closely with data engineering and data science peers to put your models into production. You bring enough breadth in ML and AI to leverage embedding-based and LLM-augmented approaches where they strengthen the graph, and you contribute meaningfully to entity resolution and identity linking work that depends on the semantic layer you define. This role reports to the Data Science Manager, Knowledge Graph & Identity. ### What You'll Do *Ontology Design & Governance* - Own the end-to-end design, development, and versioning of Samba TV's core ontologies in RDF/RDFS/OWL - defining entity classes, properties, hierarchies, and constraints that accurately model Samba's data domain at scale - Author and maintain SHACL shapes for post-load graph validation, consistency checking, and data quality enforcement - Define and document derived-attribute schemas - genre affinity, brand affinity, topic affinity, lifecycle signals, and viewing summaries - and own the logical definitions that govern how raw events become durable graph attributes - Establish ontology design standards, change management processes, and versioning practices; evaluate alignment with W3C standards and relevant industry schemas (, EIDR, DDEX, W3C PROV) - Lead ontology design reviews with product, data engineering, and data science stakeholders - articulating trade-offs between expressivity, scalability, and query performance clearly *Event-to-Ontology Derivation* - Define the aggregation and scoring logic that transforms raw TV viewership and web activity events into the durable affinities, summaries, and inferred signals that live in the graph - Co-own derivation pipeline design with data engineering - specifying transformation logic, intermediate schemas, and validation checkpoints for Databricks/Spark pipelines that feed the materialized graph substrate - Reason carefully about what belongs in the graph vs. what should remain virtualized in the data lake - balancing query performance against storage and refresh cost *Knowledge Graph Development & AI Integration* - Build and maintain production-quality knowledge graph pipelines in Python and SPARQL - well-tested, documented, and scalable to Samba's data volumes - Design and implement entity resolution and record linkage pipelines that map real-world entities (content titles, devices, audiences, advertisers) to canonical knowledge graph nodes - Develop enrichment workflows that integrate third-party data sources (metadata providers, identity vendors, web sources) into Samba's knowledge graph in a consistent, governed way - Apply embedding-based and LLM-augmented approaches to ontology mapping, entity disambiguation, and semantic similarity problems - Support content and semantic embedding pipelines that feed into the vector store and underpin GraphRAG-based AI solutions *Cross-functional Collaboration & Mentorship* - Partner with data engineering and platform teams to ensure the knowledge graph is integrated, queryable, and production-ready at scale - Collaborate with product to translate business requirements into ontological and graph data model decisions - Formally mentor Ontology Engineers and junior data scientists on semantic modeling, SHACL design patterns, and graph best practices - Lead internal technical talks and workshops on ontology, knowledge graph, and semantic web topics ### Who You Are *Must-Haves* - 5–8 years of hands-on experience in ontology engineering, semantic data modeling, or knowledge graph development - with a demonstrable track record of production ontologies at scale - Deep expertise in W3C semantic web standards: RDF, RDFS, OWL, SPARQL 1.1, and SHACL - with hands-on experience building and validating graph schemas in a production triplestore (Amazon Neptune, Stardog, GraphDB, Jena, or equivalent) - Strong Python - production-quality, well-tested code; comfortable building data pipelines and graph processing workflows - First-principles understanding of description logics, ontology design patterns, and the practical trade-offs between OWL expressivity and triplestore scalability - Hands-on experience with entity resolution, record linkage, or deduplication at scale - mapping messy, multi-source real-world data to clean ontological representations - Bachelor's degree required in Computer Science, Information Science, Computational Linguistics, Mathematics, or a related field; Master's or PhD strongly preferred - Strong communicator - able to defend ontological modeling decisions in design reviews and explain trade-offs to non-specialist stakeholders *Strongly Preferred* - Hands-on experience with Amazon Neptune or Stardog - including data virtualization (Neptune Orion or Stardog Virtual Graphs) over data lake sources - Experience designing aggregation and derivation logic that converts raw behavioral event data into durable, graph-resident derived attributes - Domain knowledge in media, entertainment, or ad tech - TV viewership (ACR/STB), digital audience modeling (device graphs, identity resolution), or ad exposure data - Familiarity with industry content and identity schemas: EIDR, VideoObject, DDEX, or equivalent - Experience with embedding models, vector databases (Milvus, Pinecone, Weaviate), and GraphRAG architectures (LangChain/LlamaIndex) - Familiarity with GNN-based approaches to knowledge graph reasoning or entity resolution a plus - Working knowledge of PySpark and Databricks for large-scale transformation pipelines