Applied Scientist, NLP​/IR​/GenAI

Join us in building Thomson Reuters' next-generation search that powers everything from direct user queries to cutting-edge GenAI applications across our product suite. We're conducting advanced research to apply modern, state-of-the-art techniques to deliver the most accurate and comprehensive search for legal and professional information.

As an Applied Scientist
, you will:

• Innovate & Deliver
  :
  Design, build, test, and deploy end-to-end AI search solutions using neural information retrieval techniques, semantic and hybrid search, and re-ranking approaches. Develop models for information retrieval, semantic search, document re-ranking, and query understanding, including dense retrieval architectures, semantic chunking models, embedding models, cross-encoders, SLM re-rankers, and transformer-based LLM-driven approaches. Work in collaboration with engineering to ensure well-managed software delivery and reliability at scale.
• Evaluate & Optimize
  :
  Develop comprehensive data and evaluation strategies for both component-level and end-to-end quality, leveraging expert human annotation and synthetic data generation. Apply robust training and evaluation methodologies to optimize retrieval quality and latency.
• Drive Technical Decisions
  :
  Independently determine appropriate retrieval architectures, indexing strategies, ranking models, data, and evaluation strategies for IR and NLP problems. Solve search relevance, ranking, and scalability challenges in a self-directed manner while contributing effectively as part of a multidisciplinary team.
• Align & Communicate
  :
  Partner closely with Engineering and Product to translate complex challenges into scalable, production-ready solutions. Engage stakeholders to deeply understand business problems and domains, shaping objectives and goals that align AI search capabilities with product needs and business objectives.
• Advance the Field
  :
  Publish at top venues (e.g., SIGIR, ECIR, NeurIPS, ACL, EMNLP, ICLR) and contribute to patents to keep our solutions cutting-edge and competitive.

• PhD in Computer Science, AI, or a related field, or a Master's with equivalent research/industry experience.
• 3+ years of hands‑on experience building and deploying modern search or RAG systems with neural retrieval methods and deep learning models for NLP.
• Strong background in information retrieval fundamentals, including indexing, query processing, ranking and relevance modelling.
• Strong programming skills (e.g., Python) and experience with modern deep learning frameworks (e.g., PyTorch, Deep Speed, Torchtune, Llama Factory).
• Proven ability to translate complex problems into innovative AI applications.
• Publications at relevant venues such as SIGIR, ECIR, NeurIPS, ACL, EMNLP, ICLR.

• Deep understanding of neural information retrieval fundamentals: BM25, hybrid search, dense retrieval (e.g., DPR, ColBERT), cross‑encoders, bi‑encoders, late interaction models.
• Hands‑on experience designing and implementing search or RAG systems: vector databases, retrieval strategies, document chunking, metadata filtering, hybrid search, re‑ranking, context optimization, and orchestration.
• Experience developing relevant datasets and evaluation frameworks.
• Solid understanding of ML and deep learning approaches for NLP.
• Solid understanding and experience with post‑training of large language models and their application to retrieval systems.

• Extensive prior work on search, question answering or RAG over large corpora and long documents, including experience with legal or enterprise search systems.
• Experience with multi‑stage or agentic retrieval architectures and query understanding for complex information needs.
• Experience building applications for the legal domain (e.g., legal search, case law retrieval, precedent finding, document review, document drafting).
• Publications at relevant venues such as SIGIR, ECIR, NeurIPS, ACL, EMNLP, ICLR.

• Hybrid Work Model: We’ve adopted a flexible hybrid working environment (2‑3 days a week in the office depending on the role) for our office‑based roles while delivering a seamless experience that is…