AI Security Researcher

Job Summary

To help mitigate security risks associated with AI systems, the full-time AI Security Researcher will design secure architectures, audit LLM-powered web agents and chatbots, and define Brave's security strategy, with a preference for candidates in a remote or hybrid setting based in London.

Key responsibilities

• Design and execute adversarial evaluations to identify failure modes in LLM-powered web agents and chatbots
• Analyze model behavior to forecast security risks in frontier models
• Develop security-focused training and fine-tuning techniques while collaborating with engineers and researchers to implement measurable security improvements

Required qualifications

• Demonstrated experience working on Secure AI and prompt injection attacks
• Strong publication record in AI security, ML security, or closely related areas
• Proficiency in Python and familiarity with ML frameworks like PyTorch or JAX
• Experience training, fine-tuning, and evaluating large language models
• A Ph.D. in computer science is highly preferred, with exceptions considered for relevant industry experience