Research Engineer, Gemini Large Scale Pretraining, DeepMind

Google•London, England

2d•$147,000 - $210,000

About The Position

At Google, research-focused Software Engineers are embedded throughout the company to set up large-scale tests and deploy promising ideas quickly and broadly. Ideas may come from internal projects or collaborations with research programs at partner universities and technical institutes. Engineers create experiments and prototype implementations, design new architectures, and work on real-world problems across areas such as artificial intelligence, data mining, natural language processing, hardware and software performance analysis, and compiler improvements. They can also remain active contributors to the wider research community through partnerships with universities and publishing papers. At Google DeepMind, interdisciplinary teams advance AI development to address complex global challenges and accelerate product innovation. The organization applies its technologies for public benefit and scientific discovery, with safety and ethics as priorities. This Research Engineer role supports Gemini large-scale pretraining, helping researchers navigate technical challenges and improving the engineering workflows and tools used across pretraining.

Requirements

Bachelor's degree in Computer Science or a related technical field, or equivalent practical experience.
2 years of experience with software engineering involving Python-based ML/scientific libraries such as JAX, or internship experience in these areas, or Open Source Software (OSS) contributions.
1 year of experience with machine learning.

Nice To Haves

Experience with large-scale system design, distributed systems, or performance engineering.
A good understanding of engineering best practices, codebase design, and testing.
Familiarity with how modern large language models are trained.
Ability to collaborate and communicate effectively with people who have different skill sets and needs and who may be distributed across different time zones.

Responsibilities

Work on long-term projects that enable new research, involve multiple team members, and require sound software engineering and design.
Respond to time-sensitive issues arising in large-scale training runs and ongoing research.
Make systematic improvements to technical workflows and tooling used across pretraining.
Apply engineering expertise to help researchers navigate technical challenges, drawing on a deep understanding of the codebase and a broad understanding of the tech stack, model architecture, and data pipeline.
Balance good engineering practices with a flexible codebase that researchers can use and extend easily.