program managementgenerative aireverse engineeringai safetythreat intelligencered teamingchild safetyllm securityjailbreaking
Вакансия из Hirify Global, списка международных tech-компаний Для мэтча и отклика нужен Plus
Пожаловаться
Зарплата и рынок
ЗП не указанаразвернуть
На похожих ролях по рынку
$114к$135кмедиана$169к
39 в выборке
На СНГ рынке: $4.9к/мес ($3.6к - $6.3к)
77
Хорошая вакансия
развернуть
Роль описана чётко и связана с быстро развивающейся сферой безопасности ИИ, но строгий офисный режим и постоянная работа с тяжёлым контентом - важные условия, которые стоит оценить заранее.
Тяжёлый контент
Только офис
Чёткая зона ответственности
Безопасность ИИ
Мэтч & Сопровод
Для мэтча с этой вакансией нужен Plus
Подключить Plus
Описание вакансии
Текст:
Hirify AI/Оригинал
TL;DR
AI Safety & Red Teaming Lead (Generative AI/Child Safety): Leading high-stakes generative AI safety engagements and multidisciplinary on-site teams with an accent on adversarial testing, child safety, operational security, and executive reporting. Focus on approving red-teaming methodologies across multimodal systems, managing severe online-harm investigations, and translating complex findings into strategic client recommendations.
Location: London, UK; full-time, 5 days per week, on-site at a strictly secured and access-controlled office.
Company
hirify.global's Alice is a trust, safety, and security company for the AI era, providing model-hardening evaluations, pre-deployment red-teaming, runtime guardrails, and drift detection for frontier model labs, enterprises, and user-generated content platforms.
What you will do
Own end-to-end delivery of high-stakes generative AI safety engagements, from scoping and methodology design through execution, quality assurance, and final reporting.
Lead and coordinate an on-site multidisciplinary team of Child Safety SMEs, reverse engineers, and AI jailbreakers.
Review, challenge, and approve adversarial testing frameworks, exploit techniques, and automated evaluation pipelines across text, image, video, multimodal, text-to-3D, and physical AI systems.
Manage client relationships, lead technical briefings, and present findings, risk assessments, and strategic recommendations to senior stakeholders.
Enforce operational security and oversee wellbeing, exposure rotation, and resilience protocols for teams handling distressing child safety material.
Requirements
At least 7 years of experience in Trust & Safety, AI red teaming, threat intelligence, or offensive security, including at least 3 years directly managing multidisciplinary technical and investigative teams.
Proven experience owning complex, high-stakes client or enterprise programs from inception through final executive delivery under tight timelines.
Strong understanding of online child exploitation, including CSAE/CSAM, grooming dynamics, and adversarial evasion tactics.
Deep working knowledge of generative AI architectures, LLM and multimodal vulnerabilities, jailbreaking techniques, reverse engineering concepts, and automated evaluation tooling.
Exceptional written and verbal English communication skills, with experience presenting technical findings, risk assessments, and policy recommendations to senior external stakeholders.
Ability to work full-time on-site 5 days per week in the London office within a secured, access-controlled environment.
Psychological resilience and emotional maturity to oversee high-severity child safety investigations and support team wellbeing during continuous exposure to distressing material.
Culture & Benefits
Work in a secured, access-controlled environment with strict operational security requirements.
Support team wellbeing through exposure rotation, resilience protocols, and psychologically safe team leadership.
Contribute to safeguarding communicative technologies and more than 3 billion users across major AI and technology platforms.
Будьте осторожны
Если работодатель просит войти в их систему, используя iCloud/Google, прислать код/пароль, запустить код/ПО, не делайте этого. Это мошенники. Обязательно жмите «Пожаловаться» или пишите в поддержку. Подробнее в гайде →