AI Temporal Knowledge Graph Update Developer

Altrix Systems

Remote · $100k – $130k / year · Full-time · Posted 4 months ago

This role is no longer accepting applications

Don't miss the next role like this

Create a free profile with your skills. We match you to new AI roles as they open, and companies hiring can contact you directly.

Create free profile, 2 minutes

Similar open roles you can apply to now:

See all open AI jobs

This role is no longer accepting applications

Don't miss the next role like this

Create a free profile with your skills. We match you to new AI roles as they open, and companies hiring can contact you directly.

Create free profile, 2 minutes

Similar open roles you can apply to now:

See all open AI jobs

About the role

Knowledge graphs encode what is true. Temporal knowledge graphs encode what was true and when — capturing that an entity's properties change over time (a company's CEO changes, a country's capital changes, a drug's regulatory status changes) and that relationships between entities have validity periods. Without temporal modeling, a knowledge graph becomes increasingly stale: the 'CEO of' edge points to the previous CEO because the graph was never updated when the transition occurred. We build knowledge intelligence systems for 18 enterprise clients. Temporal knowledge graph maintenance is the data freshness challenge: keeping the knowledge graph current as new information arrives through news feeds, regulatory filings, corporate announcements, and public data updates. The problem has two components: detecting when an existing fact has changed (change detection), and extracting the new fact with its valid-from timestamp (temporal fact extraction). You'll build the temporal knowledge graph update pipeline: change detection models that identify when new text contradicts existing graph facts, temporal fact extraction that identifies the validity period of new assertions, and the update reconciliation system that handles conflicting information from multiple sources.

Fully remote.

Responsibilities

Build change detection models identifying when new text assertions contradict or supersede existing knowledge graph facts
Implement temporal fact extraction capturing valid-from and valid-to timestamps from news and document sources
Develop update reconciliation handling conflicting assertions from multiple authoritative sources
Create knowledge graph provenance tracking linking each fact to its source and update history
Build temporal query support for knowledge graph queries with time-bounded fact retrieval
Monitor graph freshness measuring lag between real-world fact changes and corresponding graph updates

Requirements

3+ years in NLP engineering or knowledge graph development with temporal reasoning or information extraction experience
Strong Python and transformer model skills for fact extraction and change detection
Understanding of knowledge graph data models, temporal representation, and ontology update semantics
Experience handling conflicting information from multiple sources with provenance tracking
Familiarity with RDF/SPARQL, property graphs, or similar knowledge representation frameworks