pythonbashgenerative aithreat intelligencered teamingtrust and safetychild safetyadversarial testing
Вакансия из Hirify Global, списка международных tech-компаний Для мэтча и отклика нужен Plus
Пожаловаться
66
Хорошая вакансия
развернуть
У роли четкая специализация в быстро развивающейся сфере AI safety, но постоянная работа с тяжелыми материалами и пять дней в офисе в Лондоне - серьезные личные и бытовые компромиссы.
Тяжелые материалы
Только офис
Передовой AI
Четкая специализация
Мэтч & Сопровод
Для мэтча с этой вакансией нужен Plus
Подключить Plus
Описание вакансии
Текст:
Hirify AI/Оригинал
TL;DR
Child Safety Researcher (AI): Evaluating high-risk child safety scenarios across Generative AI modalities and conducting threat intelligence research with an accent on online child exploitation, adversarial testing, and product-policy mapping. Focus on designing jailbreaks and structured test suites, building Python and Bash automations for dataset analysis, and translating evolving offender tactics into actionable safety standards.
Location: Full-time, 5 days a week, on-site at the London office in a strictly secured, access-controlled environment
Company
hirify.global, operating through Alice, provides trust, safety, and security solutions for AI systems, including model evaluations, red-teaming, runtime guardrails, and drift detection.
What you will do
Review, validate, and classify complex, graphic, and sensitive child safety cases across Generative AI text, image, video, and multimodal systems.
Research online child sexual exploitation and abuse, grooming networks, and offender communities; track evolving tactics, trends, and signals.
Design and run adversarial tests, including jailbreaks, adversarial prompts, and modality-specific test approaches for text-to-3D and physical AI systems.
Map real-world threats to AI products and features and ensure evaluations reflect real-world product use.
Build scripts and local automations to analyze datasets, extract patterns, and streamline investigative workflows.
Translate investigative findings into product recommendations and safety policies in collaboration with policy, legal, and engineering teams.
Requirements
At least 5 years of specialized experience in online child safety, trust and safety investigations, CSAM/CSAE policy enforcement, or child protection threat intelligence.
Comprehensive knowledge of online child safety harms, offender methodologies, grooming dynamics, CSAM/CSAE typologies, evasion tactics, platforms, and digital communities.
Hands-on experience evaluating, red-teaming, or stress-testing Generative AI systems, including LLMs, image/video, multimodal, or emerging modalities; experience designing jailbreaks, adversarial prompts, and structured test suites.
Ability to map threat intelligence to product features and work with policy, legal, or engineering teams to draft or refine safety standards in ambiguous environments.
Hands-on scripting experience, such as Python or Bash, and experience building local automations for dataset parsing, pattern extraction, and investigative workflows.
Psychological resilience and willingness to handle continuous exposure to graphic, harsh, and distressing child safety material.
Ability to work full-time, 5 days per week, on-site in the secured London office.
Strong written and verbal English communication skills, including the ability to synthesize complex investigative findings into clear intelligence reports and actionable product recommendations.
Culture & Benefits
Work in a strictly secured, access-controlled environment focused on AI trust, safety, and security.
Contribute to safeguarding communicative technologies used by more than 3 billion people.
Support AI model labs, enterprises, and user-generated-content platforms across the AI lifecycle.
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 →