Research Engineer, Pretraining Scaling

The role

You will own critical aspects of the production pretraining pipeline, including model operations, performance optimization, and reliability. You will debug complex issues across the full stack and design experiments to improve training efficiency and model performance.

What you will do

• Own critical aspects of the production pretraining pipeline
• Debug and resolve complex issues across the full stack
• Design and run experiments to improve training efficiency and model performance
• Respond to on-call incidents during model launches
• Build and maintain production logging, monitoring dashboards, and evaluation infrastructure
• Add new capabilities to the training codebase
• Collaborate with teammates across SF and London
• Contribute to institutional knowledge by documenting systems and lessons learned

Required qualifications

• Bachelor’s degree or an equivalent combination of education, training, and/or experience
• Field of study relevant to the role as demonstrated through coursework, training, or professional experience
• Hands-on experience training large language models, or deep expertise with JAX, TPU, PyTorch, or large-scale distributed systems

Nice to have

• Previous experience training LLM’s or working extensively with JAX/TPU, PyTorch, or other ML frameworks at scale
• Contributed to open-source LLM frameworks (e.g., open_lm, llm-foundry, mesh-transformer-jax)
• Published research on model training, scaling laws, or ML systems
• Experience with production ML systems, observability tools, or evaluation infrastructure
• Background as a systems engineer, quant, or in other roles requiring both technical depth and operational excellence

Benefits

• Competitive compensation and benefits
• Optional equity donation matching
• Generous vacation and parental leave
• Flexible working hours
• Office space

Before you apply

• We do sponsor visas, though not for every role or candidate

About the company

Anthropic is a public benefit corporation headquartered in San Francisco. Its mission is to create reliable, interpretable, and steerable AI systems.