NCP-AAI practice examsFinished
NVIDIA-Certified Professional: Agentic AI
Architecting, building, deploying and governing agentic AI: reactive, deliberative and hybrid agents, ReAct and planning, memory, multi-agent orchestration with MCP and A2A, prompts, tools and error handling, evaluation and tuning, RAG and vector databases, NIM, TensorRT-LLM, Triton, NeMo Guardrails and the NeMo Agent Toolkit, Kubernetes deployment and scaling, monitoring and maintenance, safety and compliance, and human oversight.
What you get
Papers are cut from the bank and weighted like the real exam, and no question appears in two of them. Every wrong answer comes back on later days until you have it right.
What the exam covers
| Agent Architecture and Design | 15% |
| Agent Development | 15% |
| Evaluation and Tuning | 13% |
| Deployment and Scaling | 13% |
| Cognition, Planning, and Memory | 10% |
| Knowledge Integration and Data Handling | 10% |
| NVIDIA Platform Implementation | 7% |
| Run, Monitor, and Maintain | 5% |
| Safety, Ethics, and Compliance | 5% |
| Human-AI Interaction and Oversight | 5% |
Two sample questions
1. A team optimizes for correctness and latency and wants correctness to matter four times as much. How should it configure eval_metrics?
- Set both metrics' direction to maximize
- Weights of 0.8 and 0.2 for the two metrics — correct
- Use only correctness and drop latency
- Set the latency direction to maximize
Why: Each metric has a direction and a weight; weights are normalized.
2. A team's guardrails must call an external moderation API from a Colang flow. Where should the Python code live?
- In config.yml as an inline script
- In prompts.yml as a template
- In actions.py, decorated with @action() — correct
- In the kb folder as a document
Why: Custom actions are Python functions called from flows.
The real NCP-AAI exam
| Length | 120 minutes |
|---|---|
| Questions | 60-70 |
| Pass mark | Not published; NVIDIA does not publish passing scores |
| Price | $200 USD |
| Recommended before you sit it | None formally. NVIDIA recommends 1-2 years in AI/ML roles with hands-on production agentic AI work: agent development, architecture, orchestration, multi-agent frameworks, tool and model integration, evaluation, observability, deployment and guardrails. |
| Valid for | 2 years |
Exam facts as published by NVIDIA; always check the vendor’s page before you book.
Study NCP-AAI with Chiron
Chiron is invitation-only while we are in beta, and nothing is for sale yet: invited accounts use every paper, drill and course note for free.