AI Semantic Role Labeling Model Developer

Wavion Intelligence

Hybrid · $160k – $200k / 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

Semantic role labeling — determining who did what to whom, when, where, and how in a sentence — is the NLP task that connects surface syntax to event structure. A sentence like 'The regulator fined the bank $400M for compliance failures' has a predicate (fined), an agent (the regulator), a patient (the bank), an amount ($400M), and a cause (compliance failures). SRL extracts that structure, enabling downstream applications that need to reason about events, not just identify entities. We build knowledge intelligence systems for 18 enterprise clients in finance, legal, and intelligence analysis. Semantic role labeling is the event extraction layer that upgrades our knowledge graphs from entity-focused to event-focused representations. When a news article is processed, we currently extract which entities are mentioned. With SRL, we extract what happened to or between those entities — the structured event information that turns news text into actionable intelligence. You'll lead SRL model development for our domain-specific use cases: fine-tuning on financial and legal event corpora where standard PropBank-trained models fail on domain vocabulary, building the SRL serving pipeline integrated with our NLP processing stack, and developing the event-to-knowledge-graph mapping that converts SRL output into structured knowledge graph updates.

Responsibilities

Fine-tune SRL models on financial and legal event corpora where general-domain models perform poorly
Build domain-specific frame definitions for regulatory, transactional, and legal event types
Implement high-throughput SRL serving integrated with the existing NLP document processing pipeline
Develop event-to-knowledge-graph mapping converting SRL output into structured entity-relation-event triples
Create SRL quality evaluation framework benchmarking against expert-annotated domain event extractions
Handle cross-sentence event coreference tracking events that span multiple sentences in long documents

Requirements

6+ years in NLP engineering with semantic parsing, information extraction, or structured prediction experience
Deep understanding of semantic role labeling frameworks: PropBank, FrameNet, and domain-specific frame adaptation
Strong Python and PyTorch skills with span-based extraction model development and fine-tuning
Experience processing financial or legal text with domain-specific event structure
Track record integrating structured NLP outputs into knowledge representation systems