The vendor-agnostic certification for engineers who design, implement, and maintain AI safety systems. Prove your expertise across content safety, security, and operational guardrails.
AI guardrails prevent AI systems from being used in ways they shouldn't. They enforce boundaries around what models can say, do, and access — spanning content safety, security, and operational reliability.
A guardrail engineer works across all three of these pillars, designing and maintaining the systems that keep AI safe and trustworthy in production. It is a discipline that combines security engineering, ML operations, and policy expertise into a single, critical function.
Work through five modules of free, text-based training material covering every domain on the exam. Study at your own pace.
75 scored questions in 120 minutes. Multiple choice, multiple select, and scenario-based questions. You need 70% to pass.
Pass the exam and receive your Certified AI Guardrail Engineer certificate.
Start with the free training material or jump straight to the exam.