Google DeepMind is seeking a Software Engineer to contribute to groundbreaking fundamental research and deployment in large scale pre-training, specifically focusing on advancing the state of the art in artificial intelligence through research in deep learning, computer vision, and large language models.
Requirements
- A proven track record of large scale deep learning with hands-on experience with Python and neural network training (publications, open-source projects, relevant work experience, …).
- An in-depth knowledge of Transformer models and LLM training dynamics.
- Strong knowledge and experience of Python and/or C++.
- Experience with Large Language Models and notably Vision Language Models.
- Working knowledge of JAX/Pytorch or similar frameworks.
- Experience with data analysis at scale, data visualization and data ingestion pipelines for training models.
Responsibilities
- Day to day work is centered around reinforcement learning and diving into the details of data and evaluations
- Conduct modelling research: Use empirical and theoretical insights to derive novel research ideas that improve Gemini models.
- Dive deep into specific aspects of post-training (reinforcement learning, data infrastructure) to understand and improve model dynamics.
- Be deeply integrated with the Multimodal Vision team to focus specifically on scaling of Multimodal models.
Other
- 2-5 years of relevant professional experience as a Software Engineering or Technical role.
- Ability to communicate technical ideas effectively, e.g. through discussions, whiteboard sessions, written documentation.
- Bonus points for experience with distributed systems or reinforcement learning.
- The US base salary range for this full-time position is between $141,000 - $202,000 + bonus + equity + benefits.
- Application deadline: December 8, 2025
