Чётко очерченная исследовательская роль в ведущей AI-лаборатории с заметным влиянием на безопасность Gemini и внедрением исследований в продукты; требования к PhD и профильному опыту задают высокий порог входа.
?Высокий порог входа
Передовой AI
От исследований к продукту
Сильная исследовательская среда
Мэтч & Сопровод
Для мэтча с этой вакансией нужен Plus
Подключить Plus
Описание вакансии
Текст:
Hirify AI/Оригинал
TL;DR
Research Scientist (Gemini Safety and Behavior, AI): Developing safety and alignment mechanisms for Gemini foundation models across consumer interfaces, developer APIs, and search platforms with an accent on jailbreak mitigation, red and blue teaming, and agentic AI risks. Focus on designing experiments, synthetic data and evaluation pipelines, and deploying scalable safeguards for text, multimodal, and agentic systems.
Location: New York, NY, USA; Mountain View, CA, USA; or London, UK
Salary: US $207,000–$300,000 annually, plus 20% bonus target, equity, and benefits
Company
Google DeepMind is an AI research lab developing advanced artificial intelligence systems for scientific discovery and broad public benefit, with safety and ethics as core priorities.
What you will do
Investigate safety and security jailbreaks and deliver mitigations that can be deployed at scale with product teams.
Develop red- and blue-teaming methods for frontier generative AI models across text-to-text, multimodal, and agentic capabilities.
Explore data, reasoning, and algorithmic approaches to make Gemini models safe, helpful, and broadly usable.
Improve Gemini’s adversarial robustness, focusing on high-stakes abuse risks in agentic settings.
Design and execute experiments addressing known gaps and creating new safety capabilities.
Requirements
PhD in Computer Science, a related field, or equivalent practical experience.
At least 2 years of experience in large language model safety or security.
Nice to have
Experience developing and using agentic workflows for safety, behavior, and alignment.
Experience with synthetic data generation pipelines and evaluations or mitigations for non-verifiable tasks, including LLM-as-a-judge and rubric-based rewards.
Experience taking research from concept to product and collaborating on or leading applied research projects.
Strong experimental judgment regarding baselines, ablations, and experiment selection.
Publications at NeurIPS, ICLR, or ICML.
Culture & Benefits
Work in interdisciplinary, globally distributed AI research teams focused on scientific discovery and product innovation.
Benefits include equity, standard Google benefits, and a 20% bonus target.
Individual pay depends on job-related skills, experience, education, and training.
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 →