Principal Data Scientist - Safety
Core
Define, measure, and improve detection/mitigation of harmful user-generated content and violative behaviors using data science, statistics, and causal inference to establish ground truth and prevalence metrics.
Role type
Principal Data Scientist (Content Safety)
Builds
Ground truth datasets, prevalence measurement methodologies, and strategic safety insights for Product and Engineering teams.
Domain
Internet platform safety, content moderation, user-generated content
Deliverable
production ML models
Required skills
Causal inference, statistical modeling, Python, SQL, Spark, Airflow, ML/Deep Learning, first-principles reasoning, large-scale system design
Preferred skills
Experience with content platforms, UGC moderation, advanced quantitative fields
Technologies
Python, R, SQL, Hive, Spark, Airflow
Responsibilities
Establish and monitor robust prevalence measurement methodologies; Develop and validate comprehensive ground truth datasets; Conduct exploratory analysis on threat landscapes; Design and analyze experiments for moderation features; Apply causal inference to measure safety initiative impact; Partner with ML and Data Engineering teams on detection systems.
Seniority
Principal, hands-on IC with strategy & mentorship (via careerplan.io/jobs/8247229-principal-data-scientist-safety-at-roblox)
