Research Engineer/Scientist, Training Algorithms

Training algorithms, as well as the overall training recipes, are crucial components of nearly every deep neural network pipeline. The team you would join studies optimization for deep learning, hyperparameter tuning, and other methodological topics in deep learning via large scale experimental research. Our goal is to generate scientific insights and create new methods that make deep learning practitioners dramatically more effective. We are not directly tied to any single product or application, but instead conduct research intended to advance the state of the art in deep learning methods more generally.

At Google DeepMind, we’re a team of scientists, engineers, machine learning experts and more, working together to advance the state of the art in machine learning and artificial intelligence. We aim to develop widely beneficial technologies, as well as contribute to overall understanding of machine learning systems.

As a research engineer or research scientist, you will work closely with the rest of the team on cutting-edge deep learning research projects. You will design, develop, test, deploy, maintain and improve the software necessary to accomplish large-scale empirical deep learning research. Additionally, you will generate hypotheses, run experiments, and communicate research results. Finally, you will help other teams adopt the most useful insights and methods the team produces.

• Develop and maintain deep learning research infrastructure to run large scale neural network experiments
• Implement deep learning models and training algorithms using JAX and Flax
• Train deep neural networks to achieve record-breaking results on production and academic datasets
• Conduct careful and rigorous deep learning research projects within the context of the team's research agenda, and communicate the results in technical reports and presentations
• Propose new methods and research projects that align with the team's research agenda of understanding and improving the practice of deep learning

We are looking for individuals passionate about improving the practice of deep learning itself, by reducing the toil and guesswork in getting good results with neural networks. An ideal candidate would be comfortable adopting a skeptical and curious mindset, be detail-oriented, and prioritize scientific rigor. We care a great deal about what measurements we are making, why we are making them, and exactly how they are performed.

In order to set you up for success as a Research Scientist at Google DeepMind,  we look for the following skills and experience:

• MSc or PhD in machine learning or a related field, or equivalent industry experience.
• Strong software engineering skills and extensive experience using at least one deep learning framework, such as JAX or PyTorch.
• Track record of releases, publications, and/or open source projects in deep learning

In addition, the following would be an advantage: 

• Experience using JAX specifically, since it is the deep learning framework the team uses
• Research experience and accomplishments (e.g. software artifacts, scientific publications, etc,) in neural network training dynamics, training algorithms and optimizers, training recipes, or hyperparameter tuning

The US base salary range for this full-time position is between $188,000 - $230,000+ bonus + equity + benefits. Your recruiter can share more about the specific salary range for your targeted location during the hiring process.

At Google DeepMind, we value diversity of experience, knowledge, backgrounds and perspectives and harness these qualities to create extraordinary impact. We are committed to equal employment opportunity regardless of sex, race, religion or belief, ethnic or national origin, disability, age, citizenship, marital, domestic or civil partnership status, sexual orientation, gender identity, pregnancy, or related condition (including breastfeeding) 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.