Research Scientist, Reasoning

About Google DeepMind

Artificial Intelligence could be one of humanity’s most useful inventions. At Google DeepMind, a team of scientists, engineers, and machine learning experts collaborates to advance the state of the art in AI. Their technologies are developed for widespread public benefit and scientific discovery, with safety and ethics as top priorities.

Role Overview

This role focuses on building an AI scientist, targeting generalist capabilities to tackle reasoning challenges on the path to AGI. Integrated with the Gemini core effort, the position involves identifying directions to enhance model capabilities, particularly reasoning.

Key Responsibilities

• Design and implement new features and tools within Google DeepMind’s machine learning frameworks to accelerate research.
• Collaborate with researchers on experiment design and execution involving large language models.
• Design, implement, analyze, and optimize large-scale machine learning workflows.

Required Qualifications

• Degree in Computer Science, Physics, or a related quantitative field with 5+ years of relevant experience.
• Strong background in distributed systems and large-scale dataflow pipelines.
• Self-directed engineer capable of driving research ideas from conception to production in a dynamic environment.

About You

• Comfortable with all aspects of the large language model pipeline, including model training, data pipeline development, evaluation, fine-tuning, reinforcement learning training, and tooling.
• Experienced in designing and executing large-scale workflows, understanding how they fit into broader team projects.

Diversity and Inclusion

Google DeepMind values diversity of experience, knowledge, and perspectives, fostering extraordinary impact. They are committed to equal employment opportunity regardless of sex, race, religion, ethnicity, disability, age, or any other protected basis. Accommodations are available for individuals with disabilities.