Red Team Engineer, Safeguards

Red Team Engineer on Anthropic's Safeguards team who attacks the company's products and AI systems to find weaknesses before malicious actors do. The work ranges from infrastructure vulnerabilities to abuse unique to advanced AI, such as account manipulation, payment fraud and misuse of agent and tool-use features. The engineer builds multi-step attack scenarios that emulate real threat actors across the full kill chain, develops systematic methods and automated frameworks for continuous testing, and works with product, engineering and policy teams to turn findings into fixes and detection metrics.

Anthropic wants penetration testing, red teaming or AppSec experience, model jailbreaking and prompt injection testing of agentic workflows, web application security skills with tools like Burp Suite and Metasploit, and a public body of work such as CVEs or bug bounty reports. Remote-friendly in the US, with travel.