Research Scientist - Environments, London

London, United Kingdom


The pace of development of AI worldwide is limited not just by algorithms and computational resources, but also by the environments, tasks and data sets that are needed to train and test those algorithms -- so much so that how to design environments is itself an important research question. We are therefore looking for people with research backgrounds to help design and build the next generation of reinforcement learning environments at DeepMind. To succeed in this role, you will need knowledge of and experience in research within a relevant field, such as machine learning, AI, neuroscience, or cognitive psychology. You will work collaboratively with DeepMind’s research scientists on the one hand, and with a large but focussed team of expert games designers and games engineers on the other.

About us

At DeepMind we've created a unique environment for long-term ambitious research to flourish in. Our world-class interdisciplinary team combines the best techniques from deep learning, reinforcement learning and systems neuroscience to build powerful general-purpose learning algorithms.

We’ve made a number of high profile breakthroughs towards general AI and as part of Google/Alphabet we now have all the ingredients in place to make further significant progress over the coming years. Our business has a start-up entrepreneurial culture but operates within a large, high profile global company - so you will need to be flexible and ready for anything.

The Team

As part of the Research Group, the Worlds Team provides DeepMind researchers with the best training and testing environments possible, across a variety of platforms. These range from bespoke mini-games, often inspired by cognitive neuroscience and aimed at answering specific research questions, to expansive and complex first-person environments built using modern 3D games engines. The output from the Worlds Team is crucial in helping steer DeepMind’s research forward.

Projects we’ve worked on include AlphaGo, DM Lab, and DQN on Atari as well as contributions to countless publications.

The Role

Drawing on your research background, you will work with individuals in the Worlds team and Research to develop new environments and iterate on existing ones. We design and implement our learning environments carefully as it’s likely we’ll live with them for many years to come so we apply thoughtfulness to everything we do. We take robust design and implementation seriously to ensure the delivery of high quality, repeatable, customizable research environments.

You will collaborate deeply with DeepMind’s research scientists, in all areas of research, from memory to language, from physical control to vision and more. You will help researchers understand the capabilities and limitations of our current environments as well as help to design the next generation of environments. The role offers an unusual opportunity to combine creativity with technical and scientific thinking. The focus of your work will remain on the environments and tasks, but through your work you will ultimately make a verifiable, unique and lasting contribution to the development of algorithms.

Moving on to leading and managing is a possibility, although not a requirement, of the role. We’re careful to balance responsibilities such that all designers and engineers on the team, whether managers or not, contribute directly to all project that they work on. to all projects that they work on.

About You

  • Postgraduate research experience in a relevant field, such as deep learning, reinforcement learning, computer vision, language, neuroscience, cognitive or developmental psychology, safety, control, or robotics.
  • Previous interest or experience in playing with and / or creating simulation models, or board games, is an optional plus.
  • Experience with experimental design
  • A keen interest in broadening their experience to eventually encompass all of the above areas.
  • A keen interest in specializing in environments and tasks, rather than algorithms.
  • An open mind and a lifelong love of learning.
  • A keen interest in working collaboratively in a mission-driven team.An open mind and a lifelong love of learning.

Competitive salary applies.

DeepMind welcomes applications from all sections of society.  We are committed to equal employment opportunity regardless of race, religion or belief, ethnic or national origin, disability, age, citizenship, marital, domestic or civil partnership status, sexual orientation, gender identity or any other basis as protected by applicable law.  If you have a disability or additional need that requires accommodation, please do not hesitate to let us know.

Apply now


Cover Letter/Other

Application consent for DeepMind

By clicking the “I Accept” button you expressly give your consent for the collection and use of your information as described at

Your application has been successful.

Thank You

Thank you for applying to DeepMind.

You should receive an e-mail within the next few minutes acknowledging receipt of your application.

Return to Careers Page