Claude Certified Architect – Professional Concept Library

Explore concepts across all 7 exam domains. Each domain maps to the Anthropic Claude Certified Architect – Professional (CCAR-P) certification and contains task statements with atomic concepts, glossary terms, and practice questions.

38 task statements79 concepts

Integration

Design how Claude connects to the rest of the enterprise: tool/agent configuration and authn/authz, RAG pipeline and retrieval strategy design, connection-protocol selection (MCP vs. API/CLI vs. agent-to-agent), progressive discovery, and observability and accuracy-latency tradeoffs at production scale.

7 task statements·12 concepts·72 questions

19% of exam

Solution Design

Translate business problems into end-to-end Claude solution designs, select the right pattern along the augmented-LLM/workflow/agent complexity gradient, and align every architectural tradeoff to business value.

6 task statements·11 concepts·66 questions

17% of exam

Evaluation & Testing

Define measurable success criteria across accuracy, latency, cost, safety, and security; build representative eval datasets and mixed-methodology test frameworks; run A/B tests to drive iterative improvement; diagnose production failures; optimize the cost-latency-accuracy tradeoff; and monitor live systems to close the feedback loop.

6 task statements·13 concepts·60 questions

16% of exam

Governance & Safety

Design layered guardrails and defenses against prompt injection, name the risks and failure modes of LLM systems, decide where humans must stay in the loop, and meet regulatory and ethical obligations for production Claude deployments.

6 task statements·13 concepts·54 questions

14% of exam

Stakeholder & Lifecycle

Conduct structured discovery, communicate architectural decisions and tradeoffs, manage feedback loops and SLAs, document the architecture, and support the solution across its full discovery-to-iteration lifecycle.

5 task statements·10 concepts·54 questions

14% of exam

Models & Prompting

Select a Claude model against explicit quality/latency/cost tradeoffs, design system prompts and guardrails, apply the right prompting technique, and curate the context window and prompt-reuse strategy that make these designs affordable at scale.

5 task statements·12 concepts·48 questions

13% of exam

Dev Productivity

Configure Claude Code and related tooling for teams, raise developer productivity with disciplined AI-assisted workflows, and localize and resolve operational issues in Claude-powered systems.

3 task statements·8 concepts·24 questions

7% of exam

Exam Domain Weights