Principal Scientist, AI Ventures

## Principal Scientist, AI Ventures Apply remote type:
Hybrid locations:
Switzerland, Zug, Zugtime type:
Full time posted on:
Posted Todaytime left to apply:
End Date:
July 12, 2026 (22 days left to apply) job requisition :
JREQ
201374
** Principal Scientist, AI Ventures - TR Labs
** This hybrid position is based in Zug, Switzerland.
** Thomson Reuters
** are seeking a
** Principal Scientist
** to lead and architect the applied research behind
** Thomson Reuters’ AI Ventures
** incubation unit within
** TR Labs.
** In this innovative role, you will turn frontier techniques — agentic AI and multi-agent systems, orchestration and tool use, retrieval-augmented generation, and knowledge-graph-grounded reasoning — into the products the unit ships, owning the science from hypothesis through production-ready capability. This is a hands-on leadership role for someone who pairs scientific rigor with commercial judgment, knows which approaches from the literature actually work at scale, and builds at startup speed inside a corporate environment.
** About

The Role
** As
** Principal Scientist, AI Ventures
** at
** Thomson Reuters**, you will:
* Set and drive the applied-research agenda for the flagship platform and new incubation candidates — designing and executing experiments that answer critical research questions while balancing scientific rigor with business impact.
* Explore the frontier of agentic AI and multi-agent systems — orchestration, planning, tool use, memory, and evaluation — together with the LLM, RAG, and NLP methods that power autonomous and semi-autonomous workflows, testing what delivers real value in production.
* Own end-to-end experimentation and development across the research lifecycle — from data ingestion and model selection to agent design and AI-powered applications — guiding work across research, data engineering, and evaluation.
* Translate practitioner workflows and business objectives in risk, compliance, and fraud into technical requirements and testable research hypotheses.
* Partner with Engineering to shape an agentic SDLC — embedding AI agents across the software development lifecycle to accelerate experimentation, evaluation, and the path from prototype to production.
* Champion developer productivity by putting agents to work in day-to-day engineering — using coding agents and AI-assisted workflows to increase the velocity and quality of research and delivery.
* Be accountable for research deliverables across the unit’s projects, setting the standard for experimental design, reproducibility, and the evaluation of agentic systems.
* Mentor and coach scientists and engineers, raising the bar on agentic AI, LLM, and NLP best practices.
** About You
** You’re a fit for the role of
** Principal Scientist, AI Ventures
** if you have:
* Deep knowledge of agentic AI and multi-agent systems — orchestration, planning, tool use, and memory — with active experimentation across LLMs, RAG, and prompt engineering.
* Strong grounding in modern NLP methods, with experience using knowledge graphs to provide structured grounding and reasoning over complex, real-world data.
* Excellent coding skills in Python, proficiency with Git and cloud technologies, and experience writing production-quality code.
* Demonstrated ability to set a research agenda and design experiments that balance scientific rigor with business impact, translating complex, open-ended problems into successful AI applications — with the judgment to know which techniques from the literature hold up in practice.
* Experience working with real-world, messy data in commercial or industry contexts — not just clean benchmark datasets.
* Experience co-designing an agentic SDLC and evaluation processes with engineering, including agent-assisted tooling that increases developer productivity, and moving from prototype to production-ready capability at startup speed.
* Outstanding communication and analysis skills, with the ability to partner with Product, Engineering, and business stakeholders in an agile way — and to coach and mentor scientists and engineers.
* Domain familiarity with risk, compliance, fraud, or adjacent enterprise workflows — able to…