Data Scientist, Meta Superintelligence Labs (Safety)

Description

We're seeking Data Scientists to join the Safety team within MSL (Meta Superintelligence Labs). As a Data Scientist in Safety, you will establish the analytical foundations that allow us to advance personal superintelligence safely and securely. You will help us turn complex, ambiguous risks with incomplete ground truth into measurable systems across model evaluations and online monitoring. You'll work with engineering, research, product, policy, and legal to design and build measurement and mitigation strategies, quantifying trade-offs between user friction and safety risks.

Responsibilities

Measure abuse: build statistical telemetry and measurement frameworks to detect and monitor policy violations or emergent harms Scale model evaluation: design a scalable framework adapting to evolving agentic model capabilities and safety/risk landscape, grounded in statistics Evaluate and optimize safeguards: scale offline and online performance evaluation of our safeguards using human-in-the-loop and active learning Design experiments and analyses: Conduct controlled experiments and rollouts to evaluate the impact of policy or risk definition changes and safety mitigations Build intelligence and reporting: build signals, pipelines, dashboards to enable real-time monitoring driving interventions and safety roadmaps

Qualifications

Bachelor's degree in Computer Science, Computer Engineering, relevant technical field, or equivalent practical experience Bachelor's degree in Mathematics, Statistics, Computer Science, Computer Engineering, relevant technical field, or equivalent practical experience A minimum of 6 years of work experience in analytics (minimum of 4 years with a Ph.D.) Experience with data querying languages (e.g. SQL), scripting languages (e.g. Python), and/or statistical/mathematical software (e.g. R) Experience in frontier AI products or risks, and navigating online, adversarial environments in Trust & Safety or Fraud/Risk/Security domains. Model evals, threat modelling, actor telemetry, human-in-the-loop review systems don’t sound foreign to you Familiar with fast-paced, high-ambiguity, cross-functional environments – able to jump from agentic trace deep-dives to explaining risk dimensions in plain English to policy stakeholders Background in ambiguous and sparse data environments to operationalize e.g. harm prevalence measurement, causal inference, root-cause analysis– rooted in strong quantitative/statistical foundations Master's or Ph.D. Degree in a quantitative field Experience adhering to and implementing responsible, ethical AI practices (e.g., risk assessment, bias mitigation, quality and accuracy reviews) Demonstrated ability to integrate AI tools to optimize/redesign workflows and drive measurable impact (e.g., efficiency gains, quality improvements) Demonstrated ongoing AI skill development (e.g., prompt/context engineering, agent orchestration) and staying current with emerging AI technologies

Compensation: $177,000/year to $247,000/year + bonus + equity + benefits