About the Role
This position seeks an independent, experienced data scientist to design, assess, and enhance safeguards that prevent severe harms from AI systems. The role involves transforming ambiguous challenges into structured analyses, generating insights that inform product and policy decisions, and building monitoring systems to detect risks early, measure mitigation effectiveness, and balance over-blocking with under-blocking.
Responsibilities
- Assess and refine systems designed to reduce harm, including classification models and detection workflows across high-risk domains such as biosecurity, cybersecurity, and emerging threats.
- Investigate incorrect system outputs through detailed error analysis, determine root causes, and provide actionable suggestions for improving mitigation strategies.
- Design and implement monitoring systems to measure the performance of safety controls over time and across different user groups and applications.
- Detect patterns in excessive versus insufficient blocking, assess user impact, and recommend prioritized improvements.
- Analyze user feedback, complaints, and behavioral data to uncover changes in adversarial tactics and system weaknesses.
- Extend risk surveillance into new domains, such as cybersecurity threats and potential model misuse or sabotage, in collaboration with subject matter experts.
- Present findings clearly to technical and leadership teams using concise summaries, decision-focused metrics, and transparent tradeoff analysis.
Team
Part of the Safety Systems organization, embedded within the Preparedness team focused on catastrophic risk forecasting and mitigation for frontier AI models.
Other
- Background checks for applicants will be administered in accordance with applicable law.
- Qualified applicants with arrest or conviction records will be considered for employment consistent with applicable laws, including the San Francisco Fair Chance Ordinance, the Los Angeles County Fair Chance Ordinance for Employers, and the California Fair Chance Act, for US-based candidates.
- For unincorporated Los Angeles County workers: criminal history may have a direct, adverse and negative relationship with certain job duties, potentially resulting in withdrawal of a conditional offer of employment.
- Job duties require access to secure and protected information technology systems and related data security obligations.
- We are committed to providing reasonable accommodations to applicants with disabilities.
