AI Security Specialist

AI Security Specialist

In this position, you will lead offensive security testing of AI agents, including language models that access various connectors. Your main focus will be to identify vulnerabilities and potential risks before adversaries do.

• You will design and automate multi-turn attacks involving browser misuse, terminal attacks, and API exploitation.
• Create prompt-injection attack scenarios to test the security of AI agents.
• Script repeatable tests in Python or bash inside virtual machines to ensure robustness.
• Verify policy compliance and attempt policy bypasses to strengthen security measures.
• Write clear vulnerability reports documenting your findings.

For this role, we require 2+ years of experience in offensive security or adversarial machine learning. You should also have expertise in application security techniques and knowledge of language model issues. Additionally, you must have a strong track record in vulnerability reporting and understanding of privacy and financial-risk policies.