Senior Ontologist - Knowledge Graph & Identity

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 (Schema.org, 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

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

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

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

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

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, Schema.org 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