AI Discourse Structure Parsing Engineer

Wavion Intelligence

Remote · $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

Understanding that two sentences are semantically related is table stakes. Understanding that one sentence is the thesis statement of a paragraph, another is a supporting example, and a third is a counterargument that gets refuted two sentences later — that's discourse understanding, and it's one of the hardest problems in NLP that still lacks good off-the-shelf solutions. Our knowledge platform processes 80 million documents per year for intelligence analysis, research, and legal applications. The downstream tasks — argument mining, claim verification, summary generation — all perform substantially better when the model understands discourse structure, not just semantic similarity. We need someone who can build discourse parsing that works at scale on real-world messy text, not just academic benchmarks. You'll design and implement the discourse analysis pipeline: rhetorical structure theory parsing, argumentative structure extraction, and coherence modeling for long-form documents. The research literature exists; the gap is translating it into a production system that handles 80 million documents per year without breaking. You'll publish your adaptations and collaborate with our two linguists on staff who know the theory deeply.

Fully remote.

Responsibilities

Build production discourse structure parsing for rhetorical and argumentative analysis
Implement RST parsing, argument mining, and coherence modeling at document scale
Develop evaluation benchmarks on real-world documents beyond academic test sets
Optimize parsing throughput to handle 80M+ documents per year
Integrate discourse features into downstream knowledge extraction pipelines
Publish adaptations of discourse parsing research to production-scale NLP

Requirements

6+ years in NLP research or engineering with discourse analysis or argumentation mining experience
Deep knowledge of rhetorical structure theory and discourse relations
Strong PyTorch and HuggingFace skills
Experience adapting academic NLP research to production systems at scale
Publications in discourse, argumentation, or related NLP areas