Research Scientist, Provenance, DeepMind

Research Scientist, Provenance, DeepMind Overview

This role sits within DeepMind’s Strategic Initiatives work, where the team focuses on research that helps make machine learning more trustworthy, robust, and dependable. The group brings together research scientists and engineers to tackle mission-driven problems, including areas such as the provenance of synthetic media and the reliability of data. The posting highlights recent work such as SynthID and describes the team as part of a scientific community that aims to advance AI for broad public benefit while keeping safety and ethics at the center of its work.

As a Research Scientist in this area, you would use your machine learning expertise to work alongside domain specialists and other ML researchers on strategic initiatives. The role is positioned within a global organization that emphasizes interdisciplinary collaboration, scientific discovery, and building technologies that can support major product innovation and address complex challenges at scale.

What you would do

• Contribute to active research efforts in the Strategic Initiatives Program, applying your machine learning background to help solve important open problems.
• Partner with researchers and machine learning engineers to create and refine new ML approaches for strategic and meaningful challenges.
• Share research progress, conclusions, and results clearly with both technical and non-technical audiences, using written reports and verbal presentations as needed.
• Help build effective team partnerships and take part in collaborative work that supports the broader science programme’s goals.

What the role calls for

• A PhD in machine learning, or comparable research or industry experience at a similar level.
• Experience using scripting languages and pipeline tools in practical work.
• Programming ability that can be applied in research and development settings.
• A publication in a machine learning conference on a topic connected to robustness, image generation, video generation, or classification.
• Solid quantitative skills in mathematics and statistics.
• Experience applying ML methods to problems involving robust and trustworthy model deployment is desirable.
• Experience with generative AI models for audio or video is desirable.
• Experience in adversarial robustness, data influence, or watermarking is desirable.
• Strong interest in artificial intelligence.
• A track record of research work that has delivered high-quality impact.
• Ability to work in English, since the company states that English proficiency is required for all roles unless a posting says otherwise.

Team and environment

The team describes itself as collaborative, supportive, and highly ambitious. It encourages shared learning, open collaboration, and a willingness to push beyond conventional limits. The broader DeepMind environment is presented as one where interdisciplinary teams work together to advance AI responsibly and to support scientific discovery and public benefit.

Additional information from the posting

The page includes Google’s standard equal opportunity statement, noting that the company aims to build a workforce that reflects the users it serves and does not discriminate on protected characteristics. It also references applicant privacy information, accommodation support for applicants who need it, and a note for recruitment agencies that Google does not accept agency resumes or pay fees for unsolicited submissions. The posting also explains that equity grants, where applicable, are discretionary and governed by Alphabet’s equity terms and approvals.

The page does not provide a remote, hybrid, or onsite working pattern in the text shown.

Application is through Google Careers using the Apply option on the posting.

Degree Requirement: Not Specified

Visa Sponsorship May be