NCP-AIO Study Guide

This landing page centralizes module-level study notes, scenario drills, and command runbooks aligned to the official NCP-AIO exam blueprint and study guide.

Sidebar documentation-style landing page for AI Operations Professional exam prep.

Admin-only · Official NVIDIA blueprint/study guide sources · 4/4 modules published

Recommended Study Priority

Priority	Domain	Why
Tier 1	Installation and Deployment	Largest domain; sets the control-plane and runtime baseline for every later operational decision.
Tier 1	Workload Management	Execution-critical domain for mapping use cases to compute, security, routing, and workflow behavior.
Tier 1	Troubleshooting and Optimization	High-impact domain for restoring service quickly and improving workload performance under pressure.
Tier 2	Administration	Day-2 governance domain for stable user, scheduler, and artifact management.

Domain 1 - Installation and Deployment

Exam Weight: 31%

Architecture sequencing, stack installation, and initial validation for NVIDIA AI operations platforms.

Domain Overview

Deploy the platform in the correct order and verify each layer before onboarding users, projects, and workloads.

Objectives

Describe deployment architecture and sequence.
Verify hardware and software requirements.
Configure and validate hardware firmware and software.
Implement NVIDIA Management and Monitoring stack (NVIDIA BCM, NVIDIA Mission Control and NVIDIA UFM).
Deploy NVIDIA BCM toolkit by using package management, virtual machines and/or docker container.
Install and configure NVIDIA Run:ai, Slurm and Kubernetes scheduler.
Provision users, projects and quotas.
Configure and validate NVIDIA NGC private registry and NGC API key.
Install, configure and validate NVIDIA NIM and TensorRT-LLM.
Configure and validate inference backend and endpoint.
Install and configure NVIDIA Magnum IO and workload.
Install and configure NVIDIA container toolkit on worker nodes.
Install and configure DOCA services on DPU Arm by using package manager and/or containers.

Open module page Tier 1

Domain 2 - Administration

Exam Weight: 23%

Operational administration across OS, scheduler, users/quotas, and model/data lifecycle controls.

Domain Overview

Maintain a healthy operations baseline with auditable governance and repeatable maintenance workflows.

Objectives

Perform OS management and maintenance.
Perform Kubernetes and workload scheduler management.
Perform user management, role assignment and quota management.
Perform and verify data and model management.
Perform and verify NVIDIA NGC private registry and NGC API key.

Open module page Tier 2

Domain 3 - Workload Management

Exam Weight: 23%

Design and validate workload placement, routing, resource sizing, and workflow execution controls.

Domain Overview

Convert requirements into production-ready workflows with validated model paths and observable runtime state.

Objectives

Analyze use case and determine workload requirements.
Analyze use case and determine workflow and route.
Analyze use case and determine CPU, GPU and memory requirements.
Analyze use case and determine security requirements.
Configure and validate model and dataset conversion.
Configure and validate AI workflow and route.
Configure and validate AI workloads and check status.

Open module page Tier 1

Domain 4 - Troubleshooting and Optimization

Exam Weight: 23%

Troubleshoot schedulers, model/data conversion, AI workflows, and network/fabric performance bottlenecks.

Domain Overview

Use layered diagnostics to isolate failures and apply measurable optimizations across platform and workload paths.

Objectives

Troubleshoot Kubernetes and workload scheduler.
Troubleshoot model and dataset conversion.
Troubleshoot AI workflow and route.
Troubleshoot AI workloads.
Perform fabric and network diagnostics for AI workloads.
Perform AI workload optimization.

Open module page Tier 1