Research Engineer, Multi Agent Learning, DeepMind

Research Engineer, Multi Agent Learning, DeepMind Overview

This research engineering role sits within Google DeepMind in London and focuses on building the software and infrastructure that support advanced multi-agent learning research. The position is aimed at someone who can combine strong engineering practice with a research mindset, helping turn theoretical ideas into reliable systems, scalable experiments, and reproducible results. The work is centered on improving the speed, quality, and reach of research in a highly technical AI environment.

What you will do

• Design and implement new multi-agent learning algorithms and the software frameworks that support them, with an emphasis on high performance and scalability in the JAX ecosystem, including JIT compilation, autograd, and XLA.
• Develop and maintain large simulation environments and complete research pipelines that can run experiments on DeepMind’s advanced compute stack, including very large TPU pod infrastructure.
• Work in close partnership with research scientists to convert mathematical ideas, hypotheses, and experimental plans into robust code that is suitable for production use and can be reproduced reliably.
• Take technical leadership on demanding research projects, helping define engineering direction, setting standards for code quality and maintainability, and mentoring less experienced engineers.
• Improve the full research workflow from start to finish, including data preparation, model training, and analysis of outcomes, so that the team can move faster from idea to insight.

What the role requires

• A bachelor’s degree in computer science or another relevant technical field, or equivalent practical experience gained through work.
• At least five years of experience developing software in Python or C++.
• Hands-on experience with deep learning frameworks such as JAX or PyTorch.
• Experience using scientific and analytical computing tools such as numpy, pandas, and matplotlib.
• A PhD is preferred, especially if the research focus was machine learning, reinforcement learning, or multi-agent systems.
• Experience training large-scale models on accelerator hardware such as TPUs or GPUs in distributed settings is preferred.
• Experience building systems around language models, including agent-style harnesses, memory retrieval, or fine-tuning workflows, is preferred.
• A history of leading complex software initiatives, together with a strong motivation to support breakthrough research through engineering, is preferred.
• Deep expertise in building and optimizing sophisticated systems in JAX is preferred.
• Strong knowledge of multi-agent reinforcement learning, algorithmic game theory, or computational economics is preferred.
• English proficiency is required for all roles unless the posting explicitly says otherwise.
• The position is based in London, UK.

Team and environment

Google describes research-focused software engineers as being embedded across the company, where they help launch large-scale experiments and move promising ideas into practice quickly. The work can involve collaboration with internal teams as well as with universities and technical institutes around the world. Engineers in this environment contribute to experiments, prototypes, new architectures, and a wide range of technical problems spanning artificial intelligence, data mining, natural language processing, performance analysis, compiler improvement, and search-related systems.

Within Google DeepMind, the team is presented as part of a pioneering AI lab made up of interdisciplinary specialists working to advance AI for major global challenges and to support high-quality product innovation for a very large user base. The organization emphasizes public benefit, scientific discovery, safety, and ethics, and it offers opportunities to learn and grow across a global company.

Additional information

• Google states that it is an equal opportunity and affirmative action employer.
• Applicants who need workplace adjustments can request accommodation through the company’s applicants form.
• Google notes that it does not accept agency resumes and will not pay fees for unsolicited submissions.
• Equity, when granted, is discretionary and handled by Alphabet under the relevant stock plan and grant agreement; it is not guaranteed and does not affect statutory payments.

How to apply

The posting includes an application link through Google Careers. Candidates can apply directly from the job page and may also use the options to copy the link or email the listing to a friend.

Visa and sponsorship

The page does not provide any statement about visa sponsorship, work authorization, or whether this role is eligible for sponsorship.

Degree Requirement: Bachelor’s

Visa Sponsorship May be