AI Security Researcher - Secure LLMs & Web Agents

Brave; Greater London; Hybrid; GBP 120,000 - 190,000; Full time

Job summary

Brave in London is seeking an AI Security Researcher/Research Engineer to mitigate security risks in AI systems, especially LLM-powered web agents and chatbots. You design secure architectures, audit Brave’s AI agents, and help define Brave’s security direction for identifying, measuring, and reducing AI risk.

You will lead adversarial evaluations, model behavior analysis, and security-focused training. Collaboration with engineers and privacy/ML researchers will translate findings into safer

Qualifications

• Demonstrated experience working on Secure AI (prompt injection attacks and security guardrails)
• Strong publication record in AI security or related areas
• Experience shipping products and features in a fast-moving environment
• Deep familiarity with large language models and web agents
• Experience applying security principles (information flow control, access control)

Responsibilities

• Design and execute adversarial evaluations to uncover failure modes in LLM-powered web agents and chatbots (prompt injection, tool abuse, data exfiltration)
• Analyze model behavior to forecast security risks in frontier models
• Develop and prototype security-focused training and fine-tuning techniques
• Augment web agents with security principles including information flow control and least privilege
• Partner with engineers and researchers to translate findings into product changes and measurable security improvements
• Help define a forward-looking security strategy and policy for web agents to protect users

Skills

Python

PyTorch

JAX

Secure AI

LLMs

Adversarial evaluation

Information flow control

Publications

Education

Ph.D. in computer science

Tools

PyTorch

JAX

Job description

Brave in London is seeking an AI Security Researcher/Research Engineer to mitigate security risks in AI systems, especially LLM-powered web agents and chatbots. You design secure architectures, audit Brave’s AI agents, and help define Brave’s security direction for identifying, measuring, and reducing AI risk.

You will lead adversarial evaluations, model behavior analysis, and security-focused training. Collaboration with engineers and privacy/ML researchers will translate findings into safer