Ultimate Anthropic Claude Certified Architect - Foundations

This course contains the use of artificial intelligence.

Ultimate Anthropic Claude Certified Architect - Foundations prepares you for the CCAR-F exam with hands-on labs, 2 full-length practice exams, and in-depth video instruction.

Building with Claude in production takes more than knowing the API or writing a good prompt.

The Anthropic Claude Certified Architect - Foundations exam tests sound architecture and implementation decisions for real-world Claude systems. That includes agentic loops, tool design, MCP integration, Claude Code configuration, structured output, context management, reliability, human review, and production tradeoffs.

Please note: Claude certifications are currently available only to employees of Anthropic partner organizations. (It's easy to sign  your company up, though.)

This is not a trivia course. The exam is scenario-based, and this course is built the same way. You will learn how to reason through realistic production situations and choose the most appropriate architecture, configuration, or implementation pattern.

You will learn how to:

• Design agentic loops that use stop_reason, tool calls, tool results, and termination conditions correctly

• Build multi-agent systems using coordinator and subagent patterns

• Pass context between agents without losing attribution, constraints, or source metadata

• Use hooks, prerequisite gates, and structured handoff patterns when prompt guidance is not enough

• Design effective tools with clear descriptions, schemas, boundaries, and structured error responses

• Build and integrate MCP servers into Claude Code and agent workflows

• Configure Claude Code for team use with markdown files, path-specific rules, skills, slash commands, and permission modes

• Use plan mode, direct execution, iterative refinement, and CI/CD workflows appropriately

• Engineer prompts with explicit criteria, few-shot examples, structured output, JSON schemas, validation, retries, and feedback loops

• Design batch processing, multi-pass review, and multi-instance review strategies

• Manage context windows as a budget across long sessions, large codebases, documents, summaries, and handoffs

• Make reliability decisions around uncertainty, escalation, provenance, conflicting sources, and human review

The course includes 6 hands-on labs, 5 exam-style quizzes, and two full-length 60-question mock exams. The labs are technical and runnable, using Python, Claude Code, the Claude Agent SDK, the Claude API, and MCP.

The course follows the official CCAR-F exam blueprint, including:

• Agentic Architecture and Orchestration

• Tool Design and MCP Integration

• Claude Code Configuration and Workflows

• Prompt Engineering and Structured Output

• Context Management and Reliability

You will work through the same kinds of scenarios the exam emphasizes, including customer support resolution agents, multi-agent research systems, developer productivity tooling, Claude Code for CI/CD, structured data extraction, and code generation in real projects.

This course is intended for solution architects, senior engineers, developers, technical leads, and AI engineers who want to build production-grade Claude systems and prepare seriously for the CCAR-F exam.

Your instructors bring deep real-world experience in AI, cloud architecture, and software engineering.

Puria Izady is a Solutions Architect at AWS specializing in Generative AI and Machine Learning, and is the creator behind Pumping Code.

Frank Kane has taught AI and machine learning to over 1.2 million people around the world. He holds 26 patents in machine learning and spent 9 years at Amazon as a senior engineer and senior manager.

No course can guarantee a passing result. What this course gives you is a structured, blueprint-aligned path through the exam domains, hands-on experience with the core Claude architecture patterns, and the scenario practice you need to approach the exam with confidence.

By the end of this course, you will be prepared to reason through CCAR-F exam scenarios and make better architecture decisions when building real Claude-powered systems.