Knowledge Graph & Ontology Lead

rPotential•San Francisco, CA

5d•$230,000 - $285,000•Hybrid

About The Position

The Knowledge Graph & Ontology Lead designs and evolves the semantic model at the center of r.Potential’s enterprise product. The role owns the Global Labor Graph’s ontology, entity resolution, and release quality, turning fragmented external and enterprise evidence about work into a trustworthy model that supports enterprise decisions. The graph represents organizations, people, occupations, skills, tasks, decisions, AI agents, systems, and outcomes while retaining history, permissions, provenance, uncertainty, and conflicting claims. The Lead makes graph releases dependable through versioned, tested builds and evaluation of accuracy, unsupported relationships, and reasoning correctness. They partner with Product, Engineering, data engineers, data scientists, Security, and business owners to support recommendations, market intelligence, and enterprise workcharts; assess new methods in knowledge representation, graph learning, and language models; and address access controls, licensing, derived-use rights, and auditability. They also set technical direction and guide engineers and scientists through influence. The role works with a large and varied data estate, including job postings, career profiles, company records, staffing data, and public occupational, skill, and industry taxonomies. The technology stack includes Neo4j, Databricks, Spark, Python, SQL, PostgreSQL, Microsoft Azure, REST APIs, and microservices. The position is hybrid, with 2–3 days per week in the San Francisco office, and begins with demonstrating a useful result on a bounded enterprise problem before building methods that generalize across companies.

Requirements

At least 8 years of experience in data modeling, knowledge graphs, or applied machine learning.
Practical depth in ontologies, taxonomies, or semantic graphs that the candidate personally shaped and that powered real applications.
Hands-on experience with entity resolution and taxonomy crosswalks across large, inconsistent sources, including measuring and correcting match quality.
Strong hands-on ability in Python, SQL, and Spark.
Command of data engineering fundamentals, including reliable pipelines, data quality, dataset and schema versioning, and production monitoring.
Sound machine-learning and statistical reasoning, including the ability to challenge assumptions, failure modes, and evaluation, and to combine learned models with explicit structure where useful.
Systems thinking that connects an enterprise’s business problem, data, underlying representation, and application behavior, while recognizing wider effects of local improvements.
A record of turning difficult data or AI problems into products people use and explaining testing, failures, and how evidence changed the design.
Product judgment and clear communication when working with leadership, Product, Engineering, and enterprise customers on what to build and why conclusions deserve trust.
Experience leading technical initiatives through influence, including framing problems, guiding specialists, and integrating their work into a coherent result.

Nice To Haves

Familiarity with occupational and skill taxonomies such as ONET, SOC, ISCO, and ESCO, or commercial job and skill taxonomies.
Experience with graph databases such as Neo4j and with RDF, OWL, and SKOS or property-graph modeling.
Experience with graph learning, LLM-based extraction, process intelligence, or probabilistic reasoning.
Experience in a startup or high-growth environment.
A research background in knowledge representation or a related field; demonstrated technical ownership and product delivery are also valued.

Responsibilities

Design and evolve an ontology that coherently represents work, including organizations, people, occupations, skills, tasks, decisions, AI agents, systems, and outcomes, with relevant history and permissions.
Turn fragmented external and enterprise evidence into a unified model by resolving entities, aligning differing taxonomies, reconstructing work flows, and preserving provenance, uncertainty, and conflicting claims.
Own entity resolution and the quality of the Global Labor Graph.
Make graph releases dependable through versioned and tested builds and an evaluation suite measuring accuracy, unsupported relationships, and reasoning correctness.
Partner with Product and Engineering so recommendations, market intelligence, and enterprise workcharts can use the graph and explain results in its terms.
Evaluate whether changes to AI capabilities improve an entire workflow or shift bottlenecks elsewhere.
Test advances in knowledge representation, graph learning, and language models against meaningful baselines, and decide what to adopt, adapt, or set aside.
Partner with Engineering, Security, and business owners on access controls, licensing, derived-use rights, and auditability, including practices behind ISO 27001 and SOC 2.
Lead technical initiatives by setting direction, framing and reviewing engineers’ and scientists’ work, and raising standards for modeling and evaluation.
Demonstrate a useful result on a bounded enterprise problem, then develop methods that generalize across companies and improve with evidence.