Ontology Engineer

Hermes Intelligence builds proprietary intelligence infrastructure for investors, corporates, and public institutions, the systems behind high stakes decisions where public data isn't enough.

We engineer our own collection systems, custom AI pipelines, and delivery in real time on top of public and nonstandard sources. What makes the output trustworthy is the model underneath it: a precise, versioned description of the entities in a client's world and how they connect.

About the Team

Vextrum turns a client's real world exposure into a structured model that can be queried, monitored, and reported on: counterparties, assets, facilities, vessels, aircraft, events, and the relationships between them. The ontology is the spine of that system. It decides what our onboarding agent asks, what the platform can answer, and what a report is allowed to claim.

The Role

You will own that ontology end to end. That means deciding what exists, how it is named and identified, how it changes over time, and how we know it is true. You will work with the engineers who build our pipelines and our onboarding agent, and with clients' own risk, security, and strategy leads, turning what they tell you into a model that holds up.

High scope. High quality. High stakes.

This role suits a modeller who ships: someone who prefers fewer concepts done well, will say no to a change that makes the model worse, and works directly with the CTO without layers of process in between. We refuse red tape and hierarchy, but welcome people who build process and scale.

One thing to be clear about: we don't build AI wrappers or chatbots. We apply AI where it has an impact on the bottom line: resolving entities across millions of records, extracting structure from sources never meant to be read at scale, and putting the result in front of people making consequential decisions.

What You'll Build

• The core ontology: Entity types, relationships, identifiers, time, and provenance, with a clean separation between what a client declares and what we infer.
• Client models from conversation: The structure that lets our onboarding agent turn an interview into a validated slice of the ontology, including what to ask and how to resolve ambiguity.
• Source mappings: The rules that bring public and nonstandard sources into the model, including vessel and aircraft movement feeds such as AIS and ADS-B, with explicit handling of conflicts between sources.
• Entity resolution: Matching and deduplication logic that works across millions of sparse records, with confidence levels that carry through to the output.
• Governance and validation: Versioning, migration paths, constraints, and regression suites, so other engineers can extend the model without breaking it.
• Evaluation for applied LLM work: Test sets and checks for LLM assisted extraction and classification, so we know when a model is right and not only when it sounds right.

What We're Looking For

• Ontology and data modelling: Production experience with knowledge graphs or semantic models (RDF, OWL, SHACL, property graphs) and relational design, and the judgment to say which fits a given problem.
• Strong SQL & PostgreSQL: Schema design, constraints, query plans, and working directly in the database alongside the backend team.
• Python: Enough to build your own validation and mapping tooling and to review other people's.
• Entity Resolution & Provenance: Experience with messy, multi source data where every statement must be traceable to a source.
• Applied LLM Experience: Practical use of models for extraction, with a clear view of where they help and where they must be constrained.
• Domain Conversations: You can sit with a subject matter expert and leave with a precise model, without losing their meaning.
• Unwavering Ethics & Privacy: Strict adherence to legal and contractual boundaries.

Compensation & Benefits

• Competitive base salary + equity discussion for the right candidate
• Direct P&L attribution and performance rewards
• Private healthcare
• Remote flexibility within EU working hours
• Regular team offsites and Worldwide travel