Lead/Senior Research Scientist (Agentic RL)

Thomson Reuters

New York (NY)

Sur place

USD 180 000 - 250 000

Plein temps

Il y a 38 heures
Soyez parmi les premiers à postuler

Obtenez une réponse de cet employeur

Votre CV et votre lettre de motivation, adaptés exactement à ce que l’employeur recherche.

Passez les filtres ATS

Résumé du poste

Thomson Reuters Labs seeks scientists at senior and lead levels to advance fiduciary-grade AI for professional workflows. You will design benchmarks, train signals, and evaluation tools for open-weight models, working with experts in law, tax, and policy.

Expect collaboration with researchers and product teams to drive real-world adoption while publishing at top venues. You will contribute to model training, evaluation, and continual learning strategies, mentoring junior staff and turning new

Qualifications

  • PhD or equivalent practical experience in ML/CS.
  • Hands-on LLM post-training and/or evaluation experience at frontier model scale.
  • Strong publication record or work shaping LLM development.
  • Experience turning ideas into practical training or evaluation assets.
  • Excellent written and oral communication in fast-paced environments.

Responsabilités

  • Design benchmarks, training datasets, rewards, and signals to improve Thomson model family.
  • Design experiments and diagnostics to separate algorithmic effects from data and system artefacts.
  • Extend the evaluation suite to understand strengths/weaknesses and inform trade-offs.
  • Collaborate with Research Engineers and domain experts (law/tax) for training/evaluation assets.
  • Coordinate with product teams to understand adoption needs and drive integration.
  • Publish results at top ML/NLP venues and represent Thomson Reuters Labs.

Connaissances

Python
PyTorch
Research impact
Communication

Formation

PhD in machine learning, CS or related

Outils

NeMo-RL
lm-evaluation-harness

Description du poste

Job Description

Thomson Reuters Labs builds and ships custom models for professional workflows requiring fiduciary-grade AI. The Thomson LLM family was developed by training leading open-weight models on decades of authoritative Thomson Reuters content, with thousands of hours of subject-matter-expert input. Owning the model layer is how we earn trust in the most regulated, highest-stakes professional workflows. See the Thomson 1.0 Technical Report on Hugging Face for more on the latest model release.

Job Description

Thomson Reuters Labs builds and ships custom models for professional workflows requiring fiduciary-grade AI. The Thomson LLM family was developed by training leading open-weight models on decades of authoritative Thomson Reuters content, with thousands of hours of subject-matter-expert input. Owning the model layer is how we earn trust in the most regulated, highest-stakes professional workflows. See the Thomson 1.0 Technical Report on Hugging Face for more on the latest model release.
The research team trains and evaluates the Thomson models to perform at a frontier level for professional workflows. We work with a range of large open-weight models and use continual learning to improve them comprehensively, rather than pre-training from scratch or doing narrow fine-tuning. The training pipeline includes value alignment, data-centric mid-training on curated proprietary data, DPO, and reinforcement learning (RL) for product-ready tasks. The evaluation group builds the measurement side: expert-annotated legal and tax benchmarks, agentic systems, and human preference studies with subject-matter experts. We regularly publish work that comes out of our training and evaluation efforts.
We Are Hiring Scientists At Senior And Lead Level Across Three Areas Of Emphasis. In Your Application, Please Tell Us Which One Or Two Fit You Best

  • Agentic reinforcement learning: post-training for agentic behaviour (planning, long-context reasoning, tool calls, long-horizon tasks inside real agent harnesses), multi-turn RL and policy optimisation, reward modelling and training-signal design.
  • Model training: continued pre-training and post-training of open-weight models, DPO and RL, training data and reward design, forgetting and replay in continued pre-training.
  • Model evaluation: benchmarks and targeted evaluations for complex legal and professional workflows, LLM-as-a-judge and rubric design, human preference studies with legal experts, and evaluation results that inform training objectives.
Key Responsibilities
  • Design new benchmarks, training datasets, rewards and training signals to improve the Thomson model family, including for tasks where ground truth is scarce, delayed or partly subjective.
  • Design rigorous experiments and diagnostics (for example ablations) that separate algorithmic effects from data, reward and system artefacts, and carry results into production models.
  • Build and extend the evaluation suite to understand Thomson's strengths and weaknesses, and use the results to inform training objectives and technical trade-offs.
  • Work hand in hand with Research Engineers on our large-scale training and evaluation infrastructure, and with subject-matter experts in law, tax and other professional domains.
  • Coordinate with product teams and other model consumers to understand adoption requirements and drive integration.
  • Publish regularly at top-tier ML and NLP venues and represent Thomson Reuters Labs in the research community.
  • At Lead level: own a research agenda, turn emerging methods into a roadmap for our models, and mentor scientists and engineers.
Required Qualifications
  • PhD in machine learning, computer science or a related discipline, or equivalent practical experience.
  • Hands-on LLM post-training and/or evaluation experience, preferably at frontier model development scale, using Python and PyTorch.
  • A track record of research impact you can point to: publications at top-tier venues (for example NeurIPS, ICML, ICLR, ACL, EMNLP, COLM) and/or work that shaped an LLM development cycle.
  • Experience working closely with researchers and cross-functional stakeholders, including subject-matter experts, to turn ideas into practical training or evaluation assets.
  • Excellent written and oral communication, and comfort with ambiguity in a fast-paced environment.
  • For Lead level: 3+ years of research experience after the PhD, a record of leading a research agenda to published or deployed results, and of mentoring or supervising researchers.
Preferred Qualifications
  • Experience with RL post-training frameworks (NeMo-RL, verl, OpenRLHF, TRL) or evaluation frameworks (Inspect, lm-evaluation-harness).
  • Experience designing LLM-as-a-judge or rubric-based reward and evaluation pipelines and measuring their reliability.
  • Experience with Mixture-of-Experts or long-context models.
  • Experience with law, tax or other professional, non-verifiable domains, particularly grounding and citation.
  • Experience in experimental design and best practices for working with human data.
  • Familiarity with AI-assisted coding and using coding agents at scale.

Our use of AI within the recruitment process Thomson Reuters utilizes Artificial Intelligence (AI) to support parts of our global recruitment process. Unless you opt-out, our AI system will assess the information provided by you and compare it to the requirements listed for the role, and present the result to our recruitment personnel for further review. The AI system acts as a supporting tool, but there is always a human making the decision if you will be considered for the role.

In the United States, Thomson Reuters offers a comprehensive benefits package to our employees. Our benefit package includes market competitive health, dental, vision, disability, and life insurance programs, as well as a competitive 401k plan with company match. In addition, Thomson Reuters offers market leading work life benefits with competitive vacation, sick and safe paid time off, paid holidays (including two company mental health days off), parental leave, sabbatical leave. These benefits meet or exceeds the requirements of paid time off in accordance with any applicable state or municipal laws. Finally, Thomson Reuters offers the following additional benefits: optional hospital, accident and sickness insurance paid 100% by the employee; optional life and AD&D insurance paid 100% by the employee; Flexible Spending and Health Savings Accounts; fitness reimbursement; access to Employee Assistance Program; Group Legal Identity Theft Protection benefit paid 100% by employee; access to 529 Plan; commuter benefits; Adoption & Surrogacy Assistance; Tuition Reimbursement; and access to Employee Stock

Obtenez votre examen gratuit et confidentiel de votre CV.

ou faites glisser et déposez votre fichier ici.

Similar jobs

Postes similaires à comparer

Senior/Lead Research Engineer (Foundation Models)
Senior/Lead Research Engineer (Foundation Models)

Thomson Reuters • États-Unis

À distance
USD 118 000 - 220 000
Flexible vacation
Mental Health Days
401k with company match
+2
Lead/Senior Research Scientist (Agentic RL)
Lead/Senior Research Scientist (Agentic RL)

Thomson Reuters Corp. • Eagan (MN)

Sur place
USD 127 000 - 237 000
Lead/Senior Research Scientist (Agentic RL)
Lead/Senior Research Scientist (Agentic RL)

Thomson Reuters • États-Unis

À distance
USD 140 000 - 210 000
Flexible vacation
Mental health days off
Headspace access
+1
Research Manager (Frontier LLM)
Research Manager (Frontier LLM)

Thomson Reuters Corp. • Eagan (MN)

Sur place
USD 171 000 - 318 000
Hybrid Work Model
Flex My Way
Career Development
Senior/Lead Research Engineer (Foundation Models)
Senior/Lead Research Engineer (Foundation Models)

Thomson Reuters Corp. • Eagan (MN)

Sur place
USD 118 000 - 220 000
Flexible work options
Headspace app
Tuition reimbursement
+2
Lead Research Engineer – AI Evaluation & Benchmarking
Lead Research Engineer – AI Evaluation & Benchmarking

Thomson Reuters • Ann Arbor (MI)

Hybride
USD 127 000 - 237 000
Hybrid Work Model
Flex My Way
Career Development
+4
Lead Research Engineer – AI Evaluation & Benchmarking
Lead Research Engineer – AI Evaluation & Benchmarking

Thomson Reuters • Frisco (TX)

Hybride
USD 127 000 - 237 000
Hybrid work model
Flexible vacation
Mental health days
Lead Research Engineer - AI Evaluation & Benchmarking
Lead Research Engineer - AI Evaluation & Benchmarking

Thomson Reuters • New York (NY)

Hybride
USD 147 000 - 273 000
Hybrid Work Model
Mental Health Days off
Headspace access
+2
Lead Research Engineer - AI Evaluation & Benchmarking
Lead Research Engineer - AI Evaluation & Benchmarking

Thomson Reuters • Frisco (TX)

Hybride
USD 127 000 - 237 000
Hybrid Work Model
Career Development
Comprehensive Benefits
+2
Lead Research Engineer – AI Evaluation & Benchmarking
Lead Research Engineer – AI Evaluation & Benchmarking

Thomson Reuters • Ann Arbor (MI)

Hybride
USD 127 000 - 237 000
Hybrid Work Model
Flex My Way / Work-Life Balance
Career Development and Growth
+1