Job Title

Research Scientist, LLM Agents (Foundational Research) Are you a curious and open-minded individual with an interest in conducting stateoftheart foundational machine learning research? Thomson Reuters Labs is seeking Research Scientists with a passion for building complex agentbased AI systems in a datarich, complex academic environment driven by realworld problems. Foundational Research is the dedicated core machine learning research division of Thomson Reuters. We focus on research and development, particularly on advanced algorithms and training techniques for large language models (LLMs). We build a strong foundation of research capabilities across different areas and are looking for scientists who participate in designing, coding, conducting experiments, translating findings into concrete deliverables, and engaging with the academic community. Our focus areas include: LLM training: continued pretraining, instruction tuning, reinforcement learning alignment, distributed training, efficient ML techniques Posttraining techniques for planning, reasoning & complex workflows (e.g., reasoning models, LLMs + knowledge graphs, testtime compute, chainofthought pipelines, tool use & API calling, etc.) Datacentric machine learning (synthetic data, curriculum learning, learned data mixtures, etc.) Evaluation (benchmarks, humanintheloop, redteam/ adversarial testing, hallucination detection, ) Responsibilities Innovate and create new stateoftheart agent AI/LLM agent approaches at the cutting edge of AI research, solving realworld challenges using a wealth of data in agentic contexts. Participate in the entire research & model development lifecycle, brainstorming, coding, testing, and delivering highquality reports at leading international academic conferences. Collaborate with a global team of research engineers both within Thomson Reuters and at worldleading universities. Engage the wider community through seminars, lectures, conferences, and publications. Qualifications Completed or in the process of obtaining a PhD in a relevant discipline. Firstauthor publications in toptier conferences (e.g., NeurIPS, ICML, ICLR, ACL, EMNLP, NAACL) with a focus on agent systems, tool use, or multiagent coordination. Familiarity with one or more deeplearning frameworks (e.g., PyTorch, JAX, TensorFlow). Excellent communication skills to report and present research findings clearly, both orally and in writing. Curious and innovative disposition capable of devising novel, wellfoundated algorithmic solutions to relevant problems. Selfdriven attitude and ability to work with limited supervision. Comfortable working in fastpaced, agile environments, managing uncertainty and ambiguity. Preferred Qualifications Highimpact publications in toptier conferences or other influence in the research community. Experience in ML research beyond completing a PhD (e.g., supervision, industry experience, leading academic initiatives). Extensive experience with deeplearning frameworks and largescale model training. Experience working on agentbased systems, toolusing AI, or multiagent coordination in LLM contexts. Strong software and/or infrastructure engineering skills with evidence of productiongrade code contributions. Experience training largescale models over distributed nodes with cloud tools and providers such as AWS, Azure, GCP. Benefits Learning and development: onthejob coaching and learning, opportunity to work with cuttingedge methods and technologies. Plenty of data, compute, and highimpact problems: access to over 60,000TB of legal, regulatory, news, and tax data and major cloud computing platforms. Competitive compensation & benefits packages: base pay $80,000$100,000CAD (Ontario, Canada) plus annual bonus potential. Hybrid work model with flexible hybrid working environment. Flexibility & worklife balance through Flex My Way policies and flexible work arrangements. Career development and growth programs, tuition reimbursement, and awards. Comprehensive benefit plans: vacation, mentalhealth days, Headspace app access, retirement savings, employee incentive programs, and wellness resources. Social impact initiatives and paid volunteer days. Culture of inclusion, belonging, and strong values of customer obsession, winning, challenging thinking, learning fast, and working together. EEO Statement ThomsonReuters is an Equal Employment Opportunity Employer providing a drugfree workplace. We seek talented, qualified employees in all our operations worldwide, regardless of race, color, sex/gender, pregnancy, gender identity, national origin, religion, sexual orientation, disability, age, marital status, citizenship, veteran status, or any other protected classification under applicable law. We also make reasonable accommodations for qualified individuals with disabilities and religious beliefs in accordance with applicable law. #J-18808-Ljbffr