AI Coreference Resolution System Developer

Wavion Intelligence

Remote · $130k – $160k / 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

When a document says 'Apple announced its new product line' and then three sentences later says 'the company also revealed that it would', a human reader instantly knows 'the company' and 'it' refer to Apple. For a knowledge extraction system, resolving these references correctly is the difference between building a coherent knowledge graph and building a set of disconnected facts. Our NLP platform extracts structured knowledge from unstructured text across news, legal documents, regulatory filings, and scientific literature. Coreference resolution — identifying when different mentions refer to the same entity — is a systematic weakness in our extraction pipeline. We're extracting entity attributes without connecting them reliably to the entities they describe, which limits the quality of our knowledge graph. You'll lead the coreference resolution engineering effort: building systems that resolve entity mentions, event references, and pronoun chains across document types we handle, handling the domain-specific challenges each type presents (legal documents have specific reference conventions, scientific papers have complex author-result attribution), and integrating resolution into the existing extraction pipeline to improve knowledge graph density.

Responsibilities

Build neural coreference resolution models handling entity mentions, pronoun chains, and event references
Design domain adaptation for legal, scientific, and financial document coreference conventions
Implement long-document coreference handling for filings and articles spanning thousands of tokens
Integrate resolved coreference chains into the knowledge graph entity attribute extraction pipeline
Create evaluation benchmarks measuring resolution accuracy per document domain and reference type
Monitor production resolution quality and detect errors in high-volume extraction batches

Requirements

6+ years in NLP engineering with coreference resolution or information extraction experience
Deep understanding of neural coreference approaches and their domain adaptation challenges
Strong Python and PyTorch skills with HuggingFace NLP model experience
Experience building NLP pipelines that run reliably at production document volumes
Ability to evaluate and improve resolution quality using human annotation as ground truth