Data Scientist, Trust & Safety
Core
Building Replit's Trust & Safety and Anti-Abuse program to protect users and platform from AI-driven threats like phishing, scam hosting, and token farming.
Role type
Senior IC data scientist (trust & safety)
Builds
Measurement systems, risk models, anomaly-detection systems, and enforcement decision frameworks for abuse reduction.
Domain
AI-native platform security, fraud detection, and trust & safety
Deliverable
production ML models
Required skills
SQL, Python, predictive modeling, experiment design, graph analysis, causal inference, data pipeline construction, threshold selection
Preferred skills
anti-abuse system evaluation, entity resolution, progressive verification, KYC/identity providers, modern data stack (dbt, BigQuery, Snowflake)
Technologies
dbt, BigQuery, Snowflake, Fivetran, Amplitude, Mixpanel, Segment, Python, SQL
Responsibilities
Own analytical foundation for abuse metrics; build datasets and dbt models; develop and evaluate risk models; design offline evaluations and shadow-mode tests; define decision thresholds; investigate emerging abuse patterns; build monitoring for model drift; partner with Support/Legal on case review; communicate findings to technical and non-technical teams.
Seniority
Senior, hands-on IC