Research Engineer, Pretraining Scaling – London

Anthropic

Research Engineer, Pretraining Scaling – London Overview

Company Name Anthropic
Job Role Research Engineer, Pretraining Scaling – London
Qualifications Bachelor’s
Category IT Jobs
Job Type Full Time
Location London

Anthropic is seeking a dedicated Research Engineer to join their Pretraining Scaling team in London. This role involves managing and optimizing the company’s large-scale AI model training processes, ensuring that models are trained reliably, efficiently, and at scale. The successful candidate will work at the intersection of research and engineering, handling responsibilities that include performance tuning, hardware troubleshooting, experimental design, and incident response during model launches.

Key Responsibilities

  • Take charge of critical parts of the production pretraining pipeline, focusing on model operations, system performance, observability, and system reliability.
  • Diagnose and resolve complex issues across the entire training stack, including hardware errors, network problems, training dynamics, and evaluation infrastructure.
  • Design and conduct experiments to improve training efficiency, reduce step times, increase system uptime, and enhance overall model performance.
  • Respond to on-call incidents during model launches, diagnosing issues rapidly and coordinating solutions across teams.
  • Build and maintain logging systems, monitoring dashboards, and evaluation tools to support ongoing training operations.
  • Add new features to the training codebase, such as support for longer context windows or integrating novel architectures.
  • Collaborate closely with teams in San Francisco and London, including those specializing in tokens, architectures, and systems engineering.
  • Document systems, debugging approaches, and lessons learned to contribute to the team’s institutional knowledge.

Ideal Candidate Profile

  • Hands-on experience training large language models or deep expertise with frameworks like JAX, TPU, PyTorch, or other large-scale distributed ML systems.
  • Genuine enthusiasm for both research and engineering, with a preference for roles that balance both aspects equally.
  • Comfort with being on-call during launches, working extended hours, and solving complex problems under pressure.
  • Strong debugging skills across multiple layers of the training stack, with the ability to handle ambiguous issues.
  • Excellent communication skills and the ability to collaborate effectively across teams and time zones.
  • A passion for refining your craft as a research engineer and a commitment to responsible AI development.
  • Concern for the societal impacts of AI and a desire to contribute to responsible scaling efforts.

Preferred Qualifications

  • Previous experience with training large language models or extensive work with JAX/TPU, PyTorch, or similar frameworks at scale.
  • Contributions to open-source LLM frameworks or published research on model training, scaling laws, or ML systems.
  • Experience with production ML systems, observability tools, or evaluation infrastructure.
  • Background in systems engineering, quantitative analysis, or roles requiring operational excellence alongside technical depth.

What Makes This Role Unique

This position is highly operational, involving direct responsibility for ensuring smooth training of production models. It requires responsiveness to incidents, flexibility in priorities, and comfort with uncertainty. The role offers exceptional learning opportunities, working with some of the most advanced training runs in the industry, alongside top-tier researchers and engineers. The work can involve extended hours during launches, including evenings and weekends, but provides a rewarding experience for those who thrive in such environments.

Anthropic emphasizes a collaborative culture, valuing impact-driven research and open communication. The company is committed to diversity and inclusion, encouraging applicants from underrepresented groups to apply, and offers a comprehensive benefits package including competitive salary, flexible hours, generous leave, and a supportive office environment.

Applicants should be aware that visa sponsorship is available, with the company’s efforts to assist successful candidates in obtaining visas. All communication from recruiters will originate from @anthropic.com email addresses, and candidates are advised to be cautious of scams.

Join Anthropic to contribute to the development of safe, steerable AI systems that aim to benefit society, working within a research environment that values impact, collaboration, and continuous learning.


Degree Requirement: Bachelor’s

Visa Sponsorship Promising

To apply for this job please visit job-boards.greenhouse.io.