Research Engineer, Gemini Agent Safety Post Training, DeepMind

Google•Mountain View, CA
•$174,000 - $252,000

About The Position

As a Research Engineer in Gemini Agent Safety, you will architect and deploy alignment mechanisms, sandbox behaviors, and trajectory-level training to help Gemini agents—including software engineering and computer-use agents—operate safely and reliably at global scale. The role is part of the GenAI Alignment and Safety team, which safeguards the Gemini model family, from frontier reasoning models to deep research architectures. You will partner with pre-training, mid-training, post-training, and product engineering teams as agents take on higher-stakes tasks and complex workflows. The work includes advancing safety alignment algorithms and dynamic evaluations, and driving the technical execution required to deploy models across high-impact production surfaces. Google DeepMind describes itself as a pioneering AI lab with interdisciplinary teams focused on advancing AI to solve global challenges and support product innovation. It says its technologies are used for public benefit and scientific discovery, with safety and ethics as priorities. Individual pay is determined by job-related skills, experience, and relevant education or training. The stated US compensation is $174,000–$252,000, plus a 15% bonus target, equity, and benefits.

Requirements

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

Nice To Haves

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

Responsibilities

  • Design, implement, and maintain protocols to identify safety gaps in model behavior across Gemini and GenAI models, and feed findings into post-training to improve in-model and out-of-model protection.
  • Explore, adapt, and apply modeling and alignment techniques to address real-world safety and alignment challenges under tight launch timelines.
  • Advance model optimization, including the use of Supervised Fine-Tuning (SFT) and Reinforcement Learning (RL) techniques at massive scale.
  • Partner with post-training, evaluation, and product teams to align incentives, close safety gaps, and influence safety decisions without direct authority.

Benefits

  • 15% bonus target
  • Equity
  • Benefits are offered; specific benefits are not listed in the posting.
© 2026 Teal Labs, Inc
Privacy PolicyTerms of Service