Research Scientist, Deep Learning

DeepMind • Mountain View, CA

19h • $174,000 - $252,000

About The Position

Google DeepMind is looking for a Research Scientist to join the Deep Learning organization. This role is designed for a high-impact individual who balances deep theoretical intuition with a "builder" mindset. You will be responsible for advancing the frontier of machine learning by developing novel architectures and scaling training paradigms.
You are expected to navigate ambiguity, lead technical workstreams, and bridge the gap between abstract research concepts and robust, executable code. We are looking for a strong hacker—someone who thrives in the codebase and views implementation as a core part of the scientific process.

Requirements

• The Fundamentals: A Bachelors or Masters in Computer Science, Mathematics, or a related field, or equivalent practical experience. You must possess an expert-level grasp of the basics of ML (e.g., optimization, linear algebra, statistical learning theory).
• Hacker Mentality: Extensive experience with JAX, PyTorch, or TensorFlow. You should be comfortable "getting your hands dirty" in large codebases and debugging complex distributed systems.
• Proven Impact: A track record of leading research projects from ideation to deployment or publication.
• Builder Portfolio: Evidence of building significant software, tools, or models (e.g., open-source contributions, high-impact internal tools, or large-scale model releases).

Nice To Haves

• Experience with large-scale distributed training and performance profiling.
• Ability to translate "blue-sky" research ideas into concrete, incremental engineering milestones.
• Strong communication skills with the ability to influence both internal stakeholders and the broader research community.

Responsibilities

• Innovate & Execute: Design, implement, and evaluate complex deep learning models. You won’t just theorize; you will build the prototypes that prove the theory.
• Technical Leadership: Act as a primary contributor to large-scale research projects, guiding the technical direction and ensuring code quality and experimental rigor.
• Collaborative Building: Work closely with Research Engineers and cross-functional teams to integrate research breakthroughs into scalable infrastructure.
• Scientific Contribution: Publish high-quality research at top-tier venues (NeurIPS, ICML, ICLR, etc.) while maintaining a focus on tangible algorithmic improvements.
• Mentorship: Provide technical guidance to junior researchers and engineers, fostering a culture of excellence and rapid iteration.

What This Job Offers

• Job Type: Full-time
• Career Level: Mid Level