SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
xAI is seeking an experienced Program Manager for Prohibited & Regulated Content to drive priorities, measures, and system health across content integrity and enforcement programs. This role owns program standards and execution, translating legal and policy constraints into practical workflows and controls that scale across upstream detection systems and downstream operations.
You will set requirements and success criteria for detection and automation systems, partnering with Product, Engineering, and Operations teams to improve coverage, precision, and enforcement quality. Key responsibilities include identifying gaps in current frameworks and driving enhancements to policies, procedures, and workflows; owning escalations and quality assurance for complex cases; and using operational signals to harden standards and automation over time.
You will align cross-functional partners (Product, Engineering, Policy, Legal, Operations) to resolve dependencies and move critical work forward with clear accountability. You'll build and maintain documentation and operating standards that embed consistent execution, monitor regulatory developments, and track operational performance with clear recommendations for leadership.
The ideal candidate has 5+ years in content integrity, safety operations, compliance operations, or related program roles at scale. You should have working knowledge of regulated or prohibited content areas, experience translating policy/legal requirements into practical workflows, and proven ability to work with automated or model-driven enforcement systems. Strong analytical skills, cross-functional collaboration experience, and clear communication are essential. Preferred experience includes direct work on regulated content programs, setting requirements for detection/classifiers, owning escalations and QA, familiarity with regulatory developments, and exposure to integrity risks involving synthetic content or model misuse.
