Назад
1 день назад

Research Engineer (Gemini Agent Safety Post Training)

174 000 - 252 000$
Формат работы
onsite
Тип работы
fulltime
Грейд
junior
Английский
b2
Страна
US
Вакансия из списка Hirify.GlobalВакансия из Hirify Global, списка международных tech-компаний
Для мэтча и отклика нужен Plus

Мэтч & Сопровод

Для мэтча с этой вакансией нужен Plus

Описание вакансии

Текст:
/
TL;DR
Research Engineer (Gemini Agent Safety Post Training) (AI safety and alignment): Architecting and deploying alignment mechanisms, sandbox behaviors, and trajectory-level training for Gemini software engineering and computer-use agents with an accent on safe, reliable operation at global scale. Focus on designing safety protocols, applying SFT and RL at massive scale, and developing evaluations and red-teaming pipelines for behavioral edge cases and security vulnerabilities.

Location: Mountain View, CA, USA

Salary: $174,000–$252,000 USD annually, plus 15% bonus target, equity, and benefits.

Company

Google DeepMind is an AI research lab developing advanced artificial intelligence for complex global challenges, product innovation, public benefit, and scientific discovery.

What you will do

  • Architect and deploy alignment mechanisms, sandbox behaviors, and trajectory-level training for Gemini software engineering and computer-use agents.
  • Design, implement, and maintain protocols that identify safety gaps across Gemini and GenAI models and feed findings into post-training and in-model or out-of-model protections.
  • Explore, adapt, and apply modeling and alignment techniques to real-world safety challenges under tight launch timelines.
  • Advance model optimization using Supervised Fine-Tuning and Reinforcement Learning at massive scale.
  • Partner with pre-training, mid-training, post-training, evaluation, and product engineering teams to align incentives, close safety gaps, and influence safety decisions.

Requirements

  • Bachelor’s degree or equivalent practical experience.
  • 2 years of software development or ML research experience, or equivalent education credit including internships, student research, or pre-graduate work experience.
  • Experience with machine learning algorithms, deep learning research, and large language modeling.

Nice to have

  • Experience in AI safety, alignment, and robust decision-making, including SFT, RLHF, RLAIF, or online distillation.
  • Experience training and fine-tuning large models across distributed compute clusters, with familiarity with LLM serving infrastructure, telemetry logging, and continuous data flywheels.
  • Experience developing or red-teaming autonomous agents in interactive sandbox environments.
  • Expertise in automated evaluation frameworks, red-teaming pipelines, or dynamic benchmarks for identifying behavioral edge cases and security vulnerabilities.

Culture & Benefits

  • Collaboration with interdisciplinary teams across pre-training, mid-training, post-training, evaluation, and product engineering.
  • Work on safety foundations for the Gemini model family and high-impact production surfaces.
  • Benefits include equity and a 15% bonus target.

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 →