Lead Applied Scientist, Document Understanding

Thomson Reuters; Ann Arbor (MI); On-site; USD 147,600 - 274,200; Full time

Benefits offered by this job

Flexible benefits

Professional growth

Mental health days off

Retirement plan

Volunteer days

Job summary

Thomson Reuters seeks a Lead Applied Scientist for Document Understanding to own the full stack from design to production deployment. You will advance semantic chunking, document enrichment, and knowledge graph construction at scale for legal and accounting content.

Requires a PhD and 8+ years shipping NLP/IE/knowledge graph systems, with strong PyTorch and Transformers experience. You will mentor team members, partner with engineers, and influence AI strategy and roadmap.

Qualifications

• PhD in Computer Science, AI, NLP, or a related field; required.
• 8+ years of post-degree industry experience shipping document understanding or knowledge graph systems into production.
• Publications at ACL, EMNLP, ICLR, NeurIPS, SIGIR, KDD, or equivalent.
• Production Python and experience with PyTorch, Hugging Face Transformers, and DeepSpeed.

Responsibilities

• Design and deploy semantic chunking models for lengthy legal documents with adjustable granularity across use cases.
• Build document enrichment systems using legal and customer-defined taxonomies.
• Develop LLM-based knowledge graph construction pipelines that extract and link citations, entities, and legal concepts.
• Lead knowledge distillation to compress large models into latency-constrained, production-ready systems.
• Design evaluation frameworks using expert annotation and synthetic data.
• Own technical decisions on architecture, chunking strategy, classification approach, and knowledge extraction methods.

Skills

Publications in NLP

Mentoring

Education

PhD in Computer Science, AI or NLP

Tools

PyTorch

HuggingFace Transformers

DeepSpeed

Job description

Lead Applied Scientist, Document Understanding

About the Role

This role sits within the applied science function. You will own the design, development, and production deployment of document understanding systems that power Westlaw, PracticalLaw, and CoCounsel. The problems are real, the scale is large, and the expectation is shipped, reliable, measurable impact.

You will work across semantic chunking, document enrichment, knowledge graph construction, and synthetic data generation for complex legal, tax, and accounting content. Multiple product teams depend on what this function delivers.

About You

You hold a PhD in Computer Science, AI, NLP, or a related field, with 8+ years of post-degree industry experience taking NLP and document understanding systems from development to production at scale. You have hands-on depth across the full applied arc — model development, distillation, evaluation, and deployment. You publish, you mentor, and you measure success by what ships and performs in production.

What You\'ll Do

• Design and deploy semantic chunking models for lengthy, non-uniformly structured legal documents with adjustable granularity across use cases
• Build document enrichment systems using legal and customer-defined taxonomies
• Develop LLM-based knowledge graph construction pipelines that extract and link citations, entities, and legal concepts across diverse legal content
• Lead knowledge distillation efforts to compress large models into latency-constrained, production-ready SLMs
• Design evaluation frameworks — component-level and end-to-end — using expert annotation and synthetic data
• Own technical decisions on architecture, chunking strategy, classification approach, and knowledge extraction methods
• Partner with engineering on delivery, reliability, and scale across multiple product lines
• Provide technical input to senior leadership on AI strategy and roadmap
• Mentor applied scientists and ML practitioners on the team

Required Qualifications

• PhD in Computer Science, AI, NLP, or a related field – required
• 8+ years of post-degree industry experience shipping document understanding, information extraction, or knowledge graph systems into production — not research-only
• Publications at ACL, EMNLP, ICLR, NeurIPS, SIGIR, KDD, or equivalent
• Production Python and experience with PyTorch, Hugging Face Transformers, and DeepSpeed

Hands-on production depth required in

• Document layout analysis and semantic chunking beyond fixed-size or paragraph-based methods
• Hierarchical, multi-label document classification with domain-specific and customer-defined schemas
• Entity recognition and linking, relation extraction, citation parsing, and knowledge graph construction from unstructured text
• LLM-based information extraction, few-shot and multi-task learning, and post-training
• Knowledge distillation, model compression, and SLM deployment under latency constraints
• Synthetic data generation and annotation workflow design
• End-to-end evaluation framework design for document understanding

Preferred Qualifications

• Legal document understanding, legal IE, or legal AI experience
• Complex document structures: nested hierarchies, cross-references, non-uniform formatting
• Retrieval or QA systems over large document collections
• RAG and agentic workflows in enterprise settings
• Knowledge graph frameworks for legal or enterprise applications
• AzureML or AWS SageMaker

#LI-LP2

What\’s in it For You?

• Flexibility & Work-Life Balance: Flex My Way is a set of supportive workplace policies designed to help manage personal and professional responsibilities, including work from anywhere for up to 8 weeks per year, empowering employees to achieve a better work-life balance.
• Career Development and Growth: Our Grow My Way programming and skills-first approach ensures you have the tools and knowledge to grow, lead, and thrive in an AI-enabled future.
• Industry Competitive Benefits: Comprehensive benefit plans including flexible vacation, two company-wide Mental Health Days off, retirement savings, tuition reimbursement, and resources for mental, physical, and financial wellbeing.
• Culture: Globally recognized for inclusion and belonging, flexibility, work-life balance, and more. We live by our values: Obsess over our Customers, Compete to Win, Challenge (Y)our Thinking, Act Fast / Learn Fast, Stronger Together.
• Social Impact: Two paid volunteer days off annually and opportunities for pro-bono projects and ESG initiatives.
• Making a Real-World Impact: We help customers pursue justice, truth, and transparency, supporting the rule of law and trusted, unbiased information globally.

Our use of AI within the recruitment process: Thomson Reuters uses AI to support parts of our global recruitment process. Unless you opt out, the AI system will assess the information provided and present the result to our recruitment personnel for further review. A human makes the final decision on consideration for the role.

In the United States, Thomson Reuters offers a comprehensive benefits package, including health, dental, vision, disability, and life insurance, a competitive 401k plan with company match, vacation, sick and paid time off, holidays (including two mental health days), parental leave, sabbatical leave, and various optional coverages and programs. Benefits meet or exceed legal requirements where applicable.

Base pay ranges for eligible US locations: $147,600 – $274,200 USD. For Ontario, Canada: $140,000 – $175,000 CAD. Base pay is within range based on knowledge, skills, and experience. This role may be eligible for an annual bonus.

About Us

Thomson Reuters informs the way forward by combining trusted content and technology to empower professionals across legal, tax, accounting, compliance, government, and media. Reuters, part of Thomson Reuters, is a world-leading provider of trusted journalism and news. We are 26,000 employees across 70+ countries, committed to objective, accurate, fair, and transparent information. We are an Equal Employment Opportunity Employer and provide reasonable accommodations in accordance with applicable law. Learn more about accommodation requests and protecting yourself from fraudulent postings at thomsonreuters.com.