Research Engineer – Computer Use

About the role

Anthropic is looking for a Research Engineer to join its Computer Use team, which builds capabilities for the Claude model to see, interact with, and operate real software interfaces. The role blends fundamental research with product‑focused engineering, and contributions will directly improve Claude’s performance in Anthropic’s own products and those of its customers.

Key responsibilities

• Design and execute experiments that enhance Claude’s perception and agentic abilities.
• Develop robust evaluation frameworks for complex computer‑task performance.
• Build and maintain reinforcement‑learning training environments for computer use and vision.
• Create pipelines and tooling to test and validate large‑scale RL environments.
• Collaborate with model‑training and infrastructure teams to optimise production training workflows.
• Partner with product groups to translate research advances into shipped features.

Required profile

• Bachelor’s degree or equivalent experience in a relevant field.
• Professional software‑engineering experience with strong Python proficiency.
• Hands‑on experience training, fine‑tuning, or evaluating machine‑learning models.
• Demonstrated interest in the societal impact and safety of AI systems.

Required skills

• Python programming.
• Reinforcement learning, especially long‑horizon or sparse‑reward settings.
• Multimodal model training.
• Design of evaluations and benchmarks for agentic systems.
• Construction of RL environments and large‑scale ML infrastructure.

What we offer

• Competitive compensation ranging from $500 k to $850 k USD per year.
• Opportunity to work on cutting‑edge AI research that directly impacts products.
• Hybrid work arrangement in San Francisco.