AI Safety Data Scientist
The AI Safety Data Scientist will develop scalable safety solutions for AI products at Google using advanced machine learning and AI techniques. This role involves applying statistical and data science methods to analyze protection measures, identify shortcomings, and generate actionable insights to enhance security and drive business outcomes through data storytelling and automated data pipelines.
The Trust and Safety team at Google is dedicated to identifying and solving the biggest challenges to the safety and integrity of Google's products. This involves leveraging technical expertise, strong problem-solving skills, user insights, and proactive communication to protect users and partners from abuse across various Google products such as Search, Maps, Gmail, and Google Ads. Team members are strategic thinkers and team-players with a passion for upholding what is right, working globally and cross-functionally with Google engineers and product managers to combat abuse and fraud with urgency. The ultimate goal is to promote trust in Google and ensure the highest levels of user safety.
Specifically, the Scaled Protections team is responsible for safeguarding GenAI experiences across Google products for all users, including enterprise clients. This is achieved by developing and deploying safety classifiers, both server-side and on-device, to ensure the distribution of safe and policy-compliant content. The team adopts a data-driven approach to problem-solving and excels in employing cutting-edge AI solutions for enhanced safety.
Google's Trust and Safety team comprises abuse-fighting and user trust experts who work daily to make the internet a safer place. They partner with various Google teams to deliver robust solutions in areas such as malware, spam, and account hijacking. This diverse team of Analysts, Policy Specialists, Engineers, and Program Managers strives to reduce risk and combat abuse across all Google products, protecting users, advertisers, and publishers globally in over 40 languages.
Posted June 1, 2026