Platform Engineer, Intern

About the role

Metr is a nonprofit research group focused on evaluating AI capabilities and risks. We are seeking interns to join our infrastructure team to scale and improve the platform that powers our safety evaluations throughout 2026.

Key facts

What you'll do

• Develop and maintain our open-source platform used for LLM evaluations.
• Troubleshoot and optimize the performance of current evaluation processes.
• Manage and deploy cloud-based infrastructure.
• Create monitoring and observability tools for our platform.
• Write agents to automate internal workflows.
• Assist AI researchers in testing and probing advanced models.

Requirements

• Strong understanding of UNIX and Linux operating systems.
• Experience with low-level programming languages such as Rust, C, or C++.
• Knowledge of machine learning concepts, specifically transformers and embeddings.
• Proficiency in full-stack web development using TypeScript and React.
• Ability to learn independently and navigate complex technical problems.

Nice to have

• Experience with AWS services including ECS, Lambda, and IAM.
• Proficiency in database management with PostgreSQL.
• Familiarity with infrastructure-as-code tools like Pulumi, Terraform, or CDK.
• Background in Kubernetes administration.
• Prior work with LLM evaluation frameworks such as Inspect.

Skills & tools

• UNIX/Linux
• Rust, C, C++
• TypeScript, React
• LLM architecture (transformers, embeddings)

Practical notes

• Compensation is set at $150 per hour.
• This is an in-person role based in Berkeley.
• Visa sponsorship is available for candidates who do not currently possess US work authorization.
• The office provides daily catered meals and access to an on-site gym and shower facilities.
• There is a possibility for this position to convert into a full-time role.