Data Scientist, Preparedness
Core
Build, evaluate, and improve mitigation systems to detect and prevent catastrophic harms from frontier AI models, balancing over-blocking and under-blocking.
Role type
Senior IC data scientist (AI safety & risk mitigation)
Builds
Mitigation intelligence and monitoring systems for AI safety
Domain
AI safety, risk management, cybersecurity
Deliverable
production ML models
Required skills
SQL, Python, experimentation, causal thinking, observational inference, classifier evaluation, calibration, thresholding, error analysis, metrics design, dashboarding, root cause investigation, cross-functional collaboration
Preferred skills
Cybersecurity data science, threat modeling, adversarial dynamics, abuse patterns, security telemetry, detection systems in adversarial settings, trust & safety
Technologies
SQL, Python
Responsibilities
Evaluate and improve mitigation systems (classifiers, detection pipelines); Diagnose false positives/negatives via error analysis; Build monitoring frameworks to track mitigation effectiveness; Identify trends in over/under-blocking and propose interventions; Develop insights from customer feedback to detect adversarial behavior shifts; Expand risk monitoring into new areas (cybersecurity, model loss-of-control); Communicate results to technical and executive stakeholders
