We're looking for an AI Security Engineer to join our Red Team and help us push the boundaries of AI security. 
You'll run cutting-edge security assessments, develop novel testing methodologies, and work directly with enterprise clients to secure their AI systems. This role combines hands-on manual and automated red teaming, client engagement, and the development of automation that helps scale testing efforts. You'll thrive in this role if you want to be at the forefront of an emerging discipline, enjoy working on nascent problems, and like both breaking things and building processes that scale.This is a highly cross-functional position. AI security is still being defined, with best practices emerging in realtime. You'll be building the frameworks, methodologies, and tooling that scale our services while staying adaptable to rapid changes in the AI landscape. This role is ideal for someone who wants to take their traditional cybersecurity expertise and apply it to the new frontier of AI security and safety. Your focus will span several key areas:
Service Delivery & Client Engagement Deliver AI red teaming security assessments for enterprise customers.Collaborate with clients to scope projects, define testing requirements, and establish success criteriaConduct comprehensive security assessments of AI systems, including text-based LLM applications and multimodal agentic systemsExtend testing to tools, APIs, and services these AI systems connect toTest manually and creatively beyond automated coverage to find the failures that tooling missesAuthor detailed security assessment reports with actionable findings and remediation recommendationsPresent findings and strategic recommendations to technical and executive stakeholders through report readouts and Q&A sessionsTooling & Methodology DevelopmentBuild upon and improve our established processes and playbooks to scale AI red teaming service deliveryDevelop frameworks to ensure consistent, high-quality service deliveryFind tedious, repetitive tasks and automate them. You don't need to be a world-class developer—just someone who can build tools that make the team more effective. The goal is to automate routine work to free up time for targeted manual testing, not replace it.Research & InnovationDevelop novel red teaming methodologies for emerging modalities: image, video, audio, autonomous systemsStay ahead of the latest AI security threats, attack vectors, and defense mechanismsTranslate cutting-edge academic and industry research into practical testing approachesCollaborate with our research and product teams to continuously level up our methodologies
