Вакансия из Hirify Global, списка международных tech-компаний Для мэтча и отклика нужен Plus
Пожаловаться
Зарплата и рынок
развернуть
£65к-145к/год
£89к£130кмедиана£180к
По похожим ролям (50) ~ £130к/год
На СНГ рынке: $4.2к/мес ($3.6к - $5.1к)
91
Крутая вакансия
развернуть
Чётко очерченная инженерная роль в заметном государственном институте AI-безопасности даёт редкий доступ к передовым моделям, самостоятельность в исследованиях и поддержку развития; основные компромиссы - быстро меняющиеся приоритеты и сжатые сроки.
Высокий темп
Современный стек
Передовой AI
Доступ к исследованиям
?Разные уровни
Мэтч & Сопровод
Для мэтча с этой вакансией нужен Plus
Подключить Plus
Описание вакансии
Текст:
Hirify AI/Оригинал
TL;DR
Software Engineer - Red Team (AI Security): Building and maintaining systems for large-scale red-teaming, evaluation, and research automation across frontier AI models with an accent on automated jailbreaking, agent infrastructure, model serving, and secure cloud systems. Focus on designing reliable, reproducible platforms that run thousands of agents in parallel, extending evaluation tooling, and solving complex performance and security problems.
Location: London, UK, with hybrid working and options to work from government offices in Birmingham, Cardiff, Darlington, Edinburgh, Salford or Bristol; occasional remote work abroad may be available.
Annual salary: £65,000–£145,000, comprising base salary and technical allowance.
Company
The hirify.global is a UK government research organisation focused on understanding advanced AI risks and informing the safe development and deployment of frontier AI systems.
What you will do
Build and maintain large-scale red-teaming systems, including automated jailbreaking pipelines, alignment auditor agents and attacker agents.
Design infrastructure for automated research and engineering with many agents on the research platform.
Extend Inspect-based evaluation tooling to new target types.
Solve performance, reliability and security problems arising from running thousands of agents in parallel.
Support pre-release testing of frontier models, debug issues, and ensure reliable and reproducible results.
Translate research ideas into tested, documented and maintainable code while reviewing code and supporting researchers.
Requirements
Production-quality coding experience and the ability to work at a fast pace.
Interest in automating work with agents while accounting for their limitations.
Experience designing, shipping and maintaining complex technology products.
Experience improving technical standards through mentoring and feedback.
Strong written and verbal communication and effective collaboration with internal and external teams.
Experience working with a high-quality research team comprising scientists and engineers.
Python experience.
Experience building and maintaining systems on AWS or another cloud provider using infrastructure as code.
Understanding of frontier AI capabilities, limitations and risks, plus experience building infrastructure at scale with a high bar for security.
Experience serving and training models.
Successful candidates must pass a criminal record check and obtain BPSS clearance before appointment.
Nice to have
Cloud infrastructure or DevOps expertise, including AWS, Azure, Kubernetes, Terraform, CDK or Docker.
Cybersecurity, red-teaming, adversarial machine learning, penetration testing or capture-the-flag experience.
ML Ops experience, including vLLM, agent frameworks, fine-tuning or RAG systems.
Experience with large language model APIs and related tools.
Experience building, running or analysing AI evaluations, ideally with Inspect.
Understanding of AI misuse, loss-of-control risks and safeguard mechanisms.
Culture & Benefits
Mission-driven work with direct influence on frontier AI governance and deployment.
Access to pre-release frontier models, substantial compute and operational research support.
Opportunities to work with experts in national security, policy, AI research and adjacent sciences.
Early ownership of important problems, freedom to pursue research directions, and opportunities to publish and collaborate externally.
At least 25 days’ annual leave, public holidays, team breaks, volunteering leave, learning and development support, and conference funding.
Hybrid work, work-from-home equipment support, paid parental leave and Civil Service pension options.
Hiring process
Applications close on Tuesday 8 December 2026 at the end of the day, anywhere on Earth.
The process may include an initial assessment, screening call, technical assessment, behavioural interview and final interview with senior leadership.
Applications must contain truthful, factually accurate examples based on the candidate’s own experience; plagiarism or presenting AI-generated work as personal experience may lead to withdrawal.
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 →