AI Security Researcher/Research Engineer
Sign up free to see how well your resume matches this role.
What you'll do
- Design and execute adversarial evaluations to uncover failure modes in LLM-powered web agents and chatbots (e.g., prompt injection, tool abuse, data exfiltration)
- Analyze model behavior (including internal thoughts) to understand and 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, privilege control and contextual security
- Partner with engineers and (privacy, ML, and Cryptographic) researchers to translate findings into product changes, tests, and measurable security improvements
- Help define a forward-looking security strategy and policy for building and responsibly releasing web agents—advocating for user protection
What they're looking for
- Demonstrated experience working on Secure AI (prompt injection attacks and security guardrails)
- Strong publication record in AI security, ML security, or closely related areas
- A track record of shipping products and features inside a fast-moving environment
- Strong coding skills in Python and familiarity with ML frameworks like PyTorch or JAX
- Experience training, fine-tuning, and evaluating large language models
- Deep familiarity with large language models and web agents, how they work, and how they are trained
- Experience applying security principles (e.g., information flow control, access control/least privilege)
- Demonstrated success in bringing clarity and ownership to ambiguous technical problems
- A Ph.D. in computer science is highly preferred, but we may consider some exceptions with industry experience
Nice to have
- Knowledge of AI Systems (pipelines, security evaluation, etc.)
- Open source projects on security/privacy
- Experience with confidential computing and ZKP/SMPC protocols
- Ph.D. in computer science from a top-tier school
- A strong publication record in premier systems, performance, or ML
- Work experience in industry is desirable
Summarised by NextRaise from the employer’s description, which follows in full below.
Full description from employer
AI Security Researcher / Research Engineer
Remote or hybrid - London preferred but North America, EU, UK accepted
About Brave:
Brave is on a mission to protect the human right to privacy online. We’ve built a free web browser that blocks creepy ads and trackers by default, a private search engine with a truly independent index, private AI assistant, latest privacy and security features, and a private ad network (opt-in!) that directly rewards you for your attention. Already over 120 million people regularly use Brave for a faster, more private web. Millions more switch every month.
The internet is a sea of ads, hackers, and echo chambers. Big Tech makes huge profits off our data, and tells us what’s true and what’s not. Brave is fighting back. Join us!
Summary
We are looking for a researcher/research engineer to help mitigate the security risks that come with building AI systems-especially LLM-powered web agents and chatbots. In this role, you'll design and implement secure-by-design architectures, audit our LLM-powered web agent (https://brave.com/blog/ai-browsing/) and chatbot, and help set Brave's direction on how we identify, measure, and reduce security risk across AI products.
Responsibilities
- Design and execute adversarial evaluations to uncover failure modes in LLM-powered web agents and chatbots (e.g., prompt injection, tool abuse, data exfiltration)
- Analyze model behavior (including internal thoughts) to understand and 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, privilege control and contextual security
- Partner with engineers and (privacy, ML, and Cryptographic) researchers to translate findings into product changes, tests, and measurable security improvements
- Help define a forward-looking security strategy and policy for building and responsibly releasing web agents—advocating for user protection
You may be a good fit if you have
- Demonstrated experience working on Secure AI (prompt injection attacks and security guardrails)
- Strong publication record in AI security, ML security, or closely related areas
- A track record of shipping products and features inside a fast-moving environment
- Strong coding skills in Python and familiarity with ML frameworks like PyTorch or JAX
- Experience training, fine-tuning, and evaluating large language models
- Deep familiarity with large language models and web agents, how they work, and how they are trained
- Experience applying security principles (e.g., information flow control, access control/least privilege)
- Demonstrated success in bringing clarity and ownership to ambiguous technical problems
- A Ph.D. in computer science is highly preferred, but we may consider some exceptions with industry experience
Skills That Will Set You Apart from the Competition
- Knowledge of AI Systems (pipelines, security evaluation, etc.)
- Open source projects on security/privacy
- Experience with confidential computing and ZKP/SMPC protocols
- Ph.D. in computer science from a top-tier school
- A strong publication record in premier systems, performance, or ML
- Work experience in industry is desirable
Working at Brave - What we offer
- Ability to work at an exciting and well-funded international startup headquartered in San Francisco
- Cutting-edge technology, allowing you to work with state-of-the-art tools and software solutions
- Competitive compensation and plenty of room for personal growth
- Great international exposure and team atmosphere
We are looking to hire in London but would consider an exceptional candidate in the rest of Europe.
NO AGENCIES or RECRUITERS PLEASE - Unfortunately, we cannot work with third parties on these searches.
Company
Company facts come from this company's own listings. We only show what the postings themselves carry.