Open to candidates across North America, Europe.
Direct from the employer's own careers page
Who can apply
Read straight from the listing, so you know before you spend time on it.
Open to candidates across North America, Europe.
“Remote or hybrid - London preferred but North America, EU, UK accepted”
Headquartered in San Francisco with a preference for London, but accepting candidates across North America, the EU, and the UK.
- Async-first
- The listing doesn't promise async-first working.
- Contractor / B2B
- No mention of hiring international contractors or invoicing.
- Global stipend
- No home office or equipment budget mentioned.
- Annual retreat
- No company-paid get-togethers mentioned.
Read from the listing automatically. Always confirm the details with the company before you accept an offer.
About the role
In this role, you will help mitigate security risks associated with frontier AI systems, specifically LLM-powered web agents and conversational assistants. You will design and implement secure-by-design architectures, conduct adversarial evaluations (such as prompt injection and data exfiltration testing), and set Brave's technical direction for reducing security vulnerabilities in production AI.
What you'll do
- Design and execute adversarial evaluations to uncover failure modes in LLM-powered web agents and chatbots.
- Analyze model behavior and internal representations 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 ML, privacy, and cryptography researchers and engineers to translate findings into production tests and product updates.
What you bring
- Demonstrated experience working on Secure AI, including prompt injection attacks and security guardrails.
- Strong coding skills in Python and familiarity with modern ML frameworks like PyTorch or JAX.
- Experience training, fine-tuning, and evaluating large language models.
- Deep familiarity with web agents, how LLMs operate, and application of access control and security principles.
- Ph.D. in Computer Science preferred, or equivalent demonstrated experience shipping products in fast-paced environments.
About Brave
Brave is on a mission to protect online privacy with an independent browser, search engine, private AI assistant, and privacy-respecting ad network used by over 120 million people. Headquartered in San Francisco, Brave provides cutting-edge tools to fight Big Tech data tracking.
View Brave profile & open roles →