Tech Stack

Description

You'll design and build autonomous AI systems to understand and defend against advanced adversarial AI, creating evals, defensive agents, and cyberphysical interfaces. Your work directly shapes Anthropic's and the world's preparedness for advanced AI systems.

Requirements

• Strong software engineering skills, particularly in Python
• Experience building and working with LLM-based agents or autonomous systems
• Ability to solve ambiguously scoped, high-stakes problems
• Design and run experiments quickly, iterating fast
• Care deeply about AI safety and want real-world impact

Responsibilities

• Design and build autonomous AI systems with tool use and diverse environment operation
• Create evals and training environments for agent behavior understanding and shaping
• Develop defensive agents that can detect, disrupt, or outcompete adversarial AI
• Interface Claude with hardware platforms for cyberphysical risk understanding
• Translate technical findings into demonstrations for policymakers and public