Research Engineer - Gemini Pretraining

Google DeepMind is at the forefront of artificial intelligence research, working to advance AI for widespread public benefit and scientific discovery. We're seeking a Senior Research Engineer to join our team working on the cutting-edge Gemini project.

This role focuses on the critical task of inference-optimized model design for Gemini pretraining, requiring expertise in both theoretical machine learning and practical implementation. You'll be working with state-of-the-art Large Language Models (LLMs), focusing on optimizing architecture selection for fast inference while maintaining predictable quality.

The position demands a deep understanding of XLA primitives and practical experience with JAX on TPUs. You'll be involved in implementing various distillation techniques and curating datasets for model training. The role spans the entire LLM preparation stack, from pretraining to serving, requiring both technical depth and breadth.

We offer a competitive compensation package ranging from $136,000 to $300,000, plus bonus and equity. Our benefits include comprehensive medical and dental insurance, enhanced parental leave, and flexible working options. The role is based in New York City, with relocation support available.

The ideal candidate will thrive in an ambiguous environment, demonstrate flexibility in approach, and have a proven track record in machine learning and large-scale model development. You'll be joining a diverse team of scientists, engineers, and ML experts working together to push the boundaries of AI technology.

This is an exceptional opportunity to work on groundbreaking AI technology that could be one of humanity's most useful inventions. You'll be part of a team that prioritizes safety and ethics while collaborating on critical challenges in the field of artificial intelligence.