Job Description

Anthropic is hiring a Research Engineer, Computer Use to help advance Claude's ability to see, understand, and reliably operate computer interfaces.

Anthropic's Computer Use team is working on a practical challenge: teaching Claude to understand computer interfaces and use software safely and reliably.

As a Research Engineer, you'll help improve those capabilities through experimentation, reinforcement learning, evaluation, and production engineering.

The work sits between AI research and real-world product development.

You'll design experiments, test model capabilities, build training environments, and develop better ways to evaluate how well AI systems complete complex computer tasks.

Your work won't stay in a research environment. Anthropic says successful research advances can feed directly into its products and customer applications.

What You'll Do

• Run experiments to improve Claude's perception and agentic abilities.
• Build evaluations for complex computer-use tasks.
• Develop computer-use and vision reinforcement learning environments.
• Create tools for testing and validating RL environments.
• Work with model training and infrastructure teams.
• Help improve production training systems.
• Work with product teams to move research into production.

What Anthropic Is Looking For

You'll need a solid software engineering background and strong Python skills.

You should also have experience training, fine-tuning, or evaluating machine learning models. Strong communication and collaboration skills matter too.

Experience in the following areas would be especially useful:

• Computer-use or AI agent models.
• Reinforcement learning.
• Long-horizon or sparse-reward learning.
• Multimodal models.
• Agent evaluation and benchmarking.
• RL environments and simulation.
• Large-scale ML infrastructure.
• Research-to-product development.

Anthropic also wants people who take the broader impact and safety of AI seriously.

Education and Experience

A bachelor's degree or equivalent combination of education, training, and experience is listed as the minimum education requirement.

The required field of study should be relevant to the position through coursework, training, or professional experience.

Anthropic does not give a fixed number of years of experience. The requirement depends on the internal level assigned to the role.

Who This Role Could Suit

This position is particularly relevant to engineers who want to work on the next generation of AI agents and computer-use systems.

A strong candidate could come from a background in machine learning engineering, AI research, reinforcement learning, computer vision, or large-scale ML infrastructure.

You don't need to tick every box. Anthropic explicitly encourages interested candidates to apply even when they don't meet every listed qualification.

Salary and Benefits

The listed annual salary is $500,000–$850,000 USD.

Anthropic also offers a range of employee benefits, including:

• Competitive compensation.
• Vacation and parental leave.
• Flexible working hours.
• Equity donation matching.
• Collaborative office environments.

Location and Work Arrangement

The position is available in:

• San Francisco, California
• New York City, New York
• Seattle, Washington

Anthropic currently expects employees to spend at least 25% of their time in one of its offices. Some positions may require more frequent office attendance.

Visa Sponsorship

Anthropic does sponsor visas, although sponsorship is not available for every role or candidate.

The company says it makes reasonable efforts to help successful candidates obtain the necessary visa and works with an immigration lawyer during the process.

Application Close

Application deadline: Not specified. Rolling basis. Apply while the position is still open.

About Anthropic

Anthropic is an AI company focused on building reliable, interpretable, and steerable AI systems.

Its mission is to develop AI that is safe and beneficial for people and society.

The company brings together researchers, engineers, policy specialists, and business leaders. Its research spans areas such as AI safety, model capabilities, interpretability, and machine learning infrastructure.

Anthropic is headquartered in San Francisco and operates as a public benefit corporation.