NVIDIA Certified Professional: Agentic AI Exam Prep

Covers core agent architecture choices, workflow design, context/state management, tool use, prompting, and multimodal/streaming experiences. This domain emphasizes selecting patterns that meet business and technical requirements while supporting robust, adaptable agent behavior.

• Task 1.1: Select appropriate AI agent architectural patterns for business and technical requirements
• Compare architectural patterns
• Match architecture to requirements
• Identify architecture tradeoffs
• Recommend architectures for use cases
• Task 1.2: Design agent workflows that balance reactive and deliberative behavior
• Distinguish response patterns

Key references: NCP-AAI official exam guide · ExamPal shared topic tree

Covers evaluation design, metrics, dataset creation, tuning, experimentation, and monitoring for degradation and bias. This domain emphasizes aligning quality optimization with product goals such as accuracy, latency, cost, and safety.

• Task 2.1: Define evaluation strategies for AI agents and compound AI systems
• Select evaluation approaches
• Align evaluation with product goals
• Differentiate evaluation scope
• Establish success criteria
• Task 2.2: Choose and interpret task-appropriate metrics
• Select quality metrics

Key references: NCP-AAI official exam guide · ExamPal shared topic tree

Covers retrieval-augmented generation, document processing, enterprise data integration, storage and retrieval technologies, planning and reasoning, and memory architectures. This domain focuses on grounding agent behavior in reliable knowledge and supporting adaptive cognition.

• Task 3.1: Design retrieval-augmented generation and knowledge access architectures
• Select retrieval architectures
• Match retrieval to requirements
• Design grounded knowledge flows
• Optimize retrieval pipelines
• Task 3.2: Define document processing, chunking, and embedding strategies
• Choose chunking methods

Key references: NCP-AAI official exam guide · ExamPal shared topic tree

Covers implementing AI architectures with NVIDIA platform capabilities, production-ready inference and orchestration, observability, operations, and lifecycle management. The domain emphasizes reliable delivery, monitoring, and maintainability of AI systems in production.

• Task 4.1: Implement AI architectures using NVIDIA AI platform capabilities
• Identify platform components
• Map requirements to capabilities
• Select implementation patterns
• Integrate platform services
• Task 4.2: Build production-ready inference and orchestration pipelines
• Design inference paths

Key references: NCP-AAI official exam guide · ExamPal shared topic tree

Covers deployment architecture, performance and capacity optimization, security and safety controls, and governance/compliance requirements. This domain emphasizes reliable scaling, least-privilege controls, and audit-ready deployment practices.

• Task 5.1: Design deployment architectures for reliability and scale
• Select deployment patterns
• Choose resilient architectures
• Match scaling strategies
• Design for operational load
• Task 5.2: Optimize system performance and capacity
• Balance performance targets

Key references: NCP-AAI official exam guide · ExamPal shared topic tree

Covers user interaction design, transparency and trust, human oversight and escalation, and feedback-driven governance. This domain focuses on making AI-assisted workflows understandable, controllable, and safely supervised by humans.

• Task 6.1: Design user interaction patterns for AI-assisted workflows
• Select interface patterns
• Match interaction design
• Design for ambiguity resolution
• Support multimodal interaction
• Task 6.2: Improve usability, transparency, and trust
• Present evidence in outputs

Key references: NCP-AAI official exam guide · ExamPal shared topic tree