Data Scientist - AI Safety

Responsibilities

• Own safety datasets end-to-end, including collection, cleaning, sampling, labeling, quality control, versioning, and readiness for training and evaluation.
• Translate safety policy into clear, consistent labeling and evaluation criteria with policy specialists.
• Design and manage labeling processes, including sourcing, onboarding, and overseeing external contributors.
• Build evaluation workflows for production models using real-world data to track performance and surface issues.
• Develop lightweight Python and SQL pipelines and tooling to make data work faster and reproducible.
• Partner with ML engineers to define training-ready data requirements.

Requirements

• Experience as a Data Scientist or in a similar role working with real-world datasets and ML models.
• Strong understanding of dataset collection, cleaning, sampling, labeling, quality control, and evaluation.
• Proficiency with Python and SQL for analysis and practical data workflows.
• Strong critical thinking and judgment, with the ability to handle ambiguity and translate complex guidelines into consistent, scalable decisions.
• Ability to work autonomously and communicate clearly across policy, engineering, and research teams.

Nice-to-haves

• Experience in AI safety, trust and safety, content moderation, or fraud and abuse detection.
• Experience building datasets or evaluation frameworks, or working with adversarial or safety-focused machine learning.
• Experience managing human-in-the-loop labeling, dataset versioning, or ML pipeline tooling.
• Experience operationalizing guidelines with policy or legal teams.

Benefits

• Annual discretionary professional development stipend.
• Annual discretionary social travel stipend.
• Annual company offsite.
• Monthly coworking stipend for employees outside main hubs.