T

Thomson Reuters

Legal, financial data, and media services

Lead Applied Scientist - Document Understanding

Natural Language Processing & LLMsUpdated on Nov 11
Full-time

About the Job

Requirements

  • A PhD in Computer Science, Artificial Intelligence, Natural Language Processing, or a related field is required.
  • At least 8 years of industry experience building and scaling document understanding, information extraction, knowledge graph, or applied natural language processing solutions is required.
  • Demonstrated expertise across the end-to-end machine learning lifecycle, including model development, evaluation, optimization, distillation, and production deployment is required.
  • Strong Python programming skills and hands-on experience with PyTorch, Hugging Face Transformers, and DeepSpeed are required.
  • A track record of publishing research in leading conferences such as ACL, EMNLP, ICLR, NeurIPS, SIGIR, KDD, or equivalent is required.
  • The candidate must be able to mentor technical talent and drive measurable business impact through production-ready AI systems.
  • Experience translating advanced research into scalable, high-performing products and services is required.
  • The candidate must have hands-on production depth in document layout analysis and semantic chunking beyond fixed-size or paragraph-based methods.
  • The candidate must have hands-on production depth in hierarchical, multi-label document classification with domain-specific and customer-defined schemas.
  • The candidate must have hands-on production depth in entity recognition and linking, relation extraction, citation parsing, and knowledge graph construction from unstructured text.
  • The candidate must have hands-on production depth in large language model-based information extraction, few-shot and multi-task learning, and post-training.
  • The candidate must have hands-on production depth in knowledge distillation, model compression, and small language model deployment under latency constraints.
  • The candidate must have hands-on production depth in synthetic data generation and annotation workflow design.
  • The candidate must have hands-on production depth in end-to-end evaluation framework design for document understanding.

Responsibilities

  • 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 large language model-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 small language models.
  • Design component-level and end-to-end evaluation frameworks 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 artificial intelligence strategy and roadmap.
  • Mentor applied scientists and machine learning practitioners on the team.

Desired Qualifications

  • Experience with legal document understanding, legal information extraction, or legal artificial intelligence.
  • Experience with complex document structures, including nested hierarchies, cross-references, and non-uniform formatting.
  • Experience with retrieval or question-answering systems over large document collections.
  • Experience with retrieval-augmented generation and agentic workflows in enterprise settings.
  • Experience with knowledge graph frameworks for legal or enterprise applications.
  • Experience with Azure Machine Learning or AWS SageMaker.
  • Strong communication and cross-functional collaboration skills, with the ability to influence technical and business stakeholders.

About the Company

Thomson Reuters

Legal, financial data, and media services

Thomson Reuters provides intelligent information and professional services for law, tax, finance, and media. It serves law firms, corporations, government agencies, and financial institutions with platforms like legal research (Westlaw), tax software (UltraTax CS), financial market data, and news services. These products are accessed through subscription-based platforms, ensuring ongoing updates and a steady revenue stream. The company also monetizes through advertising on media platforms and data analytics sales. It distinguishes itself by offering a broad, integrated suite of industry-specific tools combined with AI, cognitive computing, and machine learning to improve accuracy, efficiency, and predictive analytics in research and data insights. Its goal is to be a trusted partner that helps professionals make informed decisions by delivering reliable information and decision-ready tools while pursuing responsible, values-driven initiatives.

Company size
10,001+
Company stage
IPO
Headquarters
Toronto, Canada
Founded
1851

Recently Posted Jobs

Explore more opportunities at Thomson Reuters

See more jobs