Overview
In this role you will steer the research agenda on agentic reinforcement learning to enable reliable, grounded LLMs for professional domains. You will collaborate with product teams, academia and engineers to design scalable training signals, evaluate in real-world contexts, and publish results. You will mentor researchers and turn ideas into deployed models, contributing to Thomson Reuters Labs' mission to advance trustworthy AI. This is a hands-on, leadership-heavy opportunity at scale.
Pay / Benefits
- Hybrid Work Model
- Flex My Way (work-life policies)
- Flexibility to work from anywhere up to 8 weeks per year
- Mental Health Days off (two company-wide)
- Headspace app access
- retirement savings, tuition reimbursement, employee incentive programs, wellbeing resources
Responsibilities
- Lead research direction and roadmap for agentic RL and reward design
- Invent dense, reliable, and cost-efficient training signals for real-world tasks
- Design controlled experiments and diagnostics to isolate algorithmic effects
- Publish regularly at top-tier venues and represent Thomson Reuters Labs
- Mentor and collaborate with scientists, engineers and research engineers
- Communicate findings through seminars, conferences, publications and technical assets
Key requirements
- PhD in machine learning, computer science or related discipline
- 3+ years of research experience after PhD in industry or academia
- Strong publication record at top-tier venues (NeurIPS, ICML, ICLR, ACL, EMNLP, NAACL, COLM) on RL for LLMs, reward modelling, reasoning, agents and tool use, or preference optimisation
- Hands-on experience post-training LLMs with RL or preference optimisation, using PyTorch (Advanced)
- Proven experience mentoring or supervising researchers, engineers or students and guiding work to published or deployed results
- Core competencies: Research leadership, Experimental judgement, Mentorship, Communication, Self-direction
- Mentorship
- Communication (written and oral)
- Self-direction and adaptability
- PyTorch (Advanced)
- RL for LLMs
- Reward modelling and preference optimisation
Lead/Senior Research Scientist β Agentic Reinforcement Learning in London employer: Thomson Reuters
Thomson Reuters is an exceptional employer, offering a dynamic work environment where innovation thrives and employees are empowered to lead the development of cutting-edge AI solutions. With a strong commitment to work-life balance, comprehensive benefits, and a culture that prioritises inclusion and professional growth, employees can expect to make a meaningful impact while advancing their careers in a globally recognised organisation. The hybrid work model and opportunities for community engagement further enhance the appeal of working in London, a vibrant hub for technology and legal expertise.