Ethics & Safety · Advanced
NeMo Guardrails and the Agent Toolkit: Programmable Safety and Governed Agents
NeMo Guardrails puts a programmable policy layer between users, models and tools — topical, safety, execution and retrieval rails defined in Colang. Pair it with the NeMo Agent Toolkit to observe and evaluate agents rather than trusting them.
- 1.5 hours lesson
- intermediate level
- NVIDIA AI Track
- Narrated presentation
What you’ll learn
- Input, dialogue, retrieval, execution and output rails as distinct enforcement points
- Colang for defining canonical forms and dialogue flows as reviewable code
- NemoGuard NIMs for content safety, topic control and jailbreak detection
- Rails cost latency and tokens — order them cheapest-check-first
- NeMo Agent Toolkit for framework-agnostic agent profiling and evaluation
- Auditing block logs for over-blocking as well as under-blocking
Inside this lesson
A narrated, slide-by-slide presentation with 1 prediction exercise, 1 knowledge check, 5 videos. You answer as you go, and the platform checks your work.