Research Engineer – Pretraining Scaling

About the role

Anthropic’s ML Performance and Scaling team builds and operates production‑grade pretrained models. As a Research Engineer you will ensure frontier models train reliably, efficiently and at scale, working across the full training stack and responding to launch‑time incidents.

Key responsibilities

• Own critical aspects of the production pretraining pipeline, including model operations, performance optimization, observability, and reliability.
• Debug and resolve complex issues across the full stack—from hardware errors and networking to training dynamics and evaluation infrastructure.
• Design and run experiments to improve training efficiency, reduce step time, increase uptime, and enhance model performance.
• Respond to on‑call incidents during model launches, diagnosing problems quickly and coordinating solutions across teams.
• Build and maintain production logging, monitoring dashboards, and evaluation infrastructure.
• Add new capabilities to the training codebase, such as long‑context support or novel architectures.
• Collaborate with teammates in London, SF and other groups (Tokens, Architectures, Systems).

Required profile

• Hands‑on experience training large language models and deep expertise with JAX, TPU, PyTorch or large‑scale distributed systems.
• Comfortable splitting time roughly 50/50 between research and engineering and enjoys both aspects.
• Willingness to be on‑call for production systems, work long days during launches and solve hard problems under pressure.
• Strong debugging skills for ambiguous, multi‑layer stack issues.
• Clear communication and effective collaboration across time zones and high‑stress incidents.
• Passion for responsible AI and the societal impacts of scaling.

Required skills

• JAX
• TPU
• PyTorch
• Large‑scale distributed systems
• Training large language models