About this role
Google DeepMind seeks a Research Scientist to improve Gemini through reinforcement learning research, new training environments and evaluations. The role involves developing ideas that can be tested at scale and transferred into production.
Responsibilities
- Develop reinforcement learning environments, evaluations and changes to the RL training recipe for Gemini.
- Explore domains where Gemini could achieve exceptional performance, such as security, hardware, performance engineering and scientific computing; design evaluations that establish current gaps.
- Invent difficult, verifiable problems and the graders used to assess them.
- Evaluate and train Gemini in these environments, examine model trajectories to understand what it learns, and deliver improvements that transfer into production.
- Conduct experiments, prototype implementations, contribute to research breakthroughs, and share or publish findings with the wider research community.
Minimum qualifications: A PhD in Computer Science, Artificial Intelligence, Machine Learning or a related technical field, or equivalent practical experience. The posting requires 1 year of experience with generative AI, large language models, natural language processing or agent-based systems, and 1 year working on modern LLM post-training—such as SFT, RLHF, DPO or PPO—model alignment or core generative model development in an industry AI lab, research institute or frontier AI organization. English proficiency is required.
Preferred qualification: Experience with reinforcement learning for LLM post-training.
The position is located in Mountain View, California. Listed US pay is $147,000–$210,000, plus a 15% bonus target, equity and benefits; individual pay depends on job-related skills, experience and relevant education or training.