Our client, an AI research and safety company, is hiring contract Safeguards Enforcement Operations Specialists. The role protects the integrity of the client’s AI systems: you’ll label data, review user interactions, enforce policy guidelines, and work across teams to keep improving safety mechanisms. This group is expanding the human review behind the client’s access and enforcement programs. Case volume is rising ahead of the automation that will eventually absorb much of it, and appeals and account compromise cases carry firm turnaround expectations. Every hire works shared review queues as a generalist and also anchors one specialty track. The AI research access and restricted-region tracks call for extra judgment.Core responsibilities (shared across Safeguards Enforcement roles)Review customer signals, user exchanges, and code to identify potential policy violations.Apply established Safeguards workflows and policy guidelines with precision and care.Work with Strategy Analysts, Policy Managers, and Threat Intelligence Investigators to spot emerging trends, highlight policy gaps, and flag content for classifier refinement.Write thorough weekly reports on operational workflows.Handle customer communications and appeals about policy enforcement.Support surge review and content labeling when the team needs it.Triage and respond to cross-functional requests with timely help and insight.Specialty Roles they're hiring for:AI Research Access & Abuse. Review access requests tied to AI research use cases and judge whether the stated research is credible.Restricted Regions & National Security. Handle access decisions and offboarding for restricted regions and national-security-sensitive cases.Account Compromise. Investigate compromised accounts and leaked credentials, and work remediation with customers.Enforcement Appeals. Review appeals of enforcement actions and recommend upholding or reversing them, with a written rationale.Account Abuse & Ban Evasion. Investigate returning accounts and clusters of related accounts, weighing multiple signals to decide on linkage and action.
