Thomson Reuters logo
Thomson ReutersPosted 1 month ago

Senior Applied Scientist, Search & Information Retrieval

$137,100–$254,700 year

On-siteToronto, Ontario, Canada or Eagan, Minnesota, United States

Full TimeSenior LevelDoctorate Or Professional DegreeEnterprise

Job Summary

Design, build, and deploy end-to-end neural search systems including dense retrieval, hybrid search, semantic chunking, and transformer-based approaches for Westlaw, PracticalLaw, and CoCounsel. Develop models for query understanding, document re-ranking, and retrieval quality optimization while driving independent technical decisions on indexing strategy and evaluation methodology. Establish evaluation frameworks using expert annotation and synthetic data generation to measure performance at scale. Contribute to published research at venues such as SIGIR, NeurIPS, and ACL, and partner with engineering teams on delivery, reliability, and cross-product scaling.

Required Qualifications

  • PhD or Master's in Computer Science, AI, NLP, or a related field
  • 5+ years of post-degree industry experience shipping search, retrieval, or RAG systems into production — not research-only experience
  • Publications at SIGIR, ECIR, NeurIPS, ACL, EMNLP, ICLR, or equivalent
  • Production Python and experience with PyTorch, DeepSpeed, Torchtune, or LlamaFactory
  • Hands-on production depth required in: Neural IR fundamentals: BM25, hybrid search, dense retrieval (DPR, ColBERT), bi-encoders, cross-encoders, late interaction models
  • Search and RAG system design: vector databases, retrieval strategies, document chunking, metadata filtering, re-ranking, context optimisation, and orchestration
  • Evaluation framework design for retrieval quality at component and system level
  • Post-training of large language models and their application to retrieval systems
  • Deep learning and NLP fundamentals

Desired Qualifications

  • Search, QA, or RAG over large corpora and long documents, including legal or enterprise search
  • Multi-stage or agentic retrieval architectures and query understanding for complex information needs
  • Legal domain applications: case law retrieval, precedent finding, document review
  • AzureML or AWS SageMaker

Hiring someone like this?

Get your role in front of qualified candidates on Sorce.

Get started

Apply to this job in one click with Sorce

Apply on Sorce