Data Scientist - AI Safety
Core
Own safety datasets end-to-end and build evaluation workflows to assess AI model performance in production.
Role type
hands-on IC data scientist (AI safety & evaluation)
Builds
production-ready safety datasets and evaluation pipelines for AI models
Domain
AI safety, content moderation, and model evaluation
Deliverable
production ML models
Required skills
dataset collection, cleaning, sampling, labelling, quality control, evaluation, Python, SQL, critical thinking, autonomous decision-making
Preferred skills
AI safety/trust & safety experience, human-in-the-loop labelling management, ML pipeline tooling, policy operationalization
Technologies
Python, SQL
Responsibilities
Own safety datasets end-to-end (collection, cleaning, labelling, QC, versioning); Translate safety policy into labelling and evaluation criteria; Design and manage labelling processes including external contributors; Build evaluation workflows for models in production; Develop lightweight Python/SQL pipelines for data work
Seniority
Mid-level, hands-on IC
