VP, Regulated AI Models - Research & Development

Refinitiv; Zug; Hybrid; CHF 180.000 - 260.000; Vollzeit

Benefits dieser Stelle

Hybrid work model

Career development

Competitive benefits

Zusammenfassung

Thomson Reuters Labs is seeking a senior leader to own the model-development roadmap end to end for professional workflows, collaborating with product, infrastructure, and SME teams. You will guide data strategy, training, evaluation, and release decisions for Thomson LLMs used across CoCounsel, Westlaw, and Practical Law.

The role requires leadership experience, a track record shipping models, and the ability to balance safety, cost, and performance in regulated environments.

Qualifikationen

• Led a production LLM program end to end (data strategy through training, evaluation, and release) for real users.
• Strong command of post-training methods (SFT, preference optimization, RL) and distributed training.
• Track record of shipped models, open-source contributions, or notable publications.

Aufgaben

• Own the model-development roadmap end to end for multiple product lines.
• Set and oversee an online reinforcement-learning pipeline in collaboration with SMEs.
• Ensure rigorous data strategy and evaluation tied to measurable outcomes.
• Enable new product capabilities built on Thomson LLMs with cost/quality advantages.
• Build and retain a senior bench of scientists and engineers, including leadership talent.
• Partner across product, infrastructure, and editorial/SME teams to align priorities.

Kenntnisse

Program ownership

Technical fluency

Leadership

Delegation with accountability

Strategic vision

Jobbeschreibung

Thomson Reuters Labs is seeking a senior leader to own the model-development roadmap end to end for professional workflows, collaborating with product, infrastructure, and SME teams. You will guide data strategy, training, evaluation, and release decisions for Thomson LLMs used across CoCounsel, Westlaw, and Practical Law.

The role requires leadership experience, a track record shipping models, and the ability to balance safety, cost, and performance in regulated environments.