Thomson Reuters logo for Research Intern - Agentic AI / NLP / ML job on DevsUnite

Research Intern - Agentic AI / NLP / ML

Thomson Reuters

Actively HiringInternship
Apply
Location

London, England, UK

Experience

0-2 years

Role

Research Intern - Agentic AI / NLP / ML

Salary

Competitive

Share this job

Know someone who'd be a great fit?

Job Description

Role Overview

Thomson Reuters Labs in London seeks a Research Intern for a six-month program focused on Agentic AI, NLP, and ML. You will own a research project aligned with TR products, collaborating with global experts to design experiments, explore novel approaches, and disseminate findings via publications.

Responsibilities

  • Drive research: lead experiments, design evaluations, analyze results
  • Innovate: explore emerging technologies advancing TR's portfolio
  • Collaborate & disseminate: work with research scientists, present internally, publish externally

Requirements

  • PhD in Computer Science, Computational Linguistics, or related field (enrolled or graduated)
  • At least 1 research publication
  • Strong NLP/ML methodology understanding, including LLMs and Agentic Workflows
  • Proven ability to design, implement, evaluate ML experiments
  • Proficiency in Python, PyTorch, TensorFlow, Hugging Face
  • Experience with unstructured data, preprocessing, data-driven problem solving
  • Familiarity with Git (GitHub/GitLab)
  • Exposure to cloud platforms (AWS, Azure, or Google Cloud)

Nice to Have

  • Strong communication skills for complex ideas
  • Prior research internship or industry experience

Benefits

  • Hybrid work model with flexible arrangements
  • Work from anywhere up to 8 weeks/year
  • Competitive benefits: flexible vacation, mental health days, Headspace, retirement, tuition reimbursement
  • Two paid volunteer days, ESG initiatives

Required Skills

PythonPyTorchTensorFlowHugging FaceGit

Ready to apply for Research Intern - Agentic AI / NLP / ML?

See how you stack up at Thomson Reuters before you apply, or go straight to the application.

Apply Anyway