Research Engineer, Autonomous Agents

Snapshot

Artificial Intelligence could be one of humanity’s most useful inventions. 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 artificial intelligence. We use our technologies for widespread public benefit and scientific discovery, and collaborate with others on critical challenges, ensuring safety and ethics are the highest priority.

The Role

The Autonomous Assistants team will conduct research into methods, models, and tools which will seek to imitate—and eventually replicate—behaviours and capabilities which make humans helpful and collaborative, and able to support one another. The team will have a strong focus on application and evaluation-driven research, whereby fundamental questions about autonomy, responsibility, and intelligent reasoning and understanding, are pragmatically studied under the lens of whether and how they improve the ability of our systems to practically help us in real-world tasks.

Key responsibilities:

• Design and implementation of use-case-focussed technology prototypes

• Design and implementation of (often human-driven) evaluation protocols

• Partner with research scientists to design and/or implement candidate technical solutions to practical challenges surrounding the assistant technology use-case(s) being focussed on by the team

• Keep up with technical developments with regard to tooling, frameworks, and other technologies surrounding: the training and fine-tuning of large models; the collection, processing, and storage of data; the design , implementation, and running of shared evaluation exercises; LLMs (including multi-modality); tool-use; reasoning, memory, and self-improvement; etc.

About You

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

• A strong academic record, ideally at MS/MSc/MEng level or above

• Ideally, industry experience involving engineering work pertaining to neural networks, or ML methods in general — i.e. experience with the workflow of a machine learning project, from idea prototyping to analysis and debugging

• Strong knowledge of Python, and strong software engineering skills (writing clear documentation, learning new frameworks and APIs, organizing codebases)

• Strong knowledge of machine learning, working knowledge of statistics

• Experience working with accelerators like GPUs and TPUs

• Experience with large-scale data processing frameworks (e.g., Apache Beam, Spark) or distributed training infrastructure.

In addition, the following would be an advantage:

• Experience with LLMs, in particular with fine-tuning and associated methods (RLHF, preference optimization, etc)

• Experience with data collection/annotation and evaluation/benchmarking, including benchmark design

• Experience in large-scale training, with a focus on compute-optimal configuration, hardware profiling, and training stability at scale.

• Experience with the production and dissemination of research artefacts (papers, technical reports, research framework documentation)

• Experience with UX/UI, or at least the ability to partner with specialist engineers to deliver useable software prototypes

• A passion for developing working prototypes of ML-powered technology, especially in the assistant technology space

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.

Applications close Monday 12th January at 17:00 GMT.