Course Overview
Ultimate Anthropic Claude Certified Architect – Foundations prepares learners for the CCAR-F exam with hands-on labs, two full-length practice exams, and detailed video instruction. The course covers essential architecture and implementation decisions for production Claude systems, including agentic loops, tool design, MCP integration, Claude Code configuration, structured output, context management, reliability, human review, and production tradeoffs.
This technical architect-level course is designed 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.
Key Takeaways
- Pass the Anthropic Claude Certified Architect – Foundations exam (CCAR-F).
- Design agentic loops, subagents, sessions, hooks, and handoff patterns for production Claude systems.
- Build MCP tools and servers with clear schemas, scoped access, structured errors, and reliable tool routing.
- Configure Claude Code for team workflows using CLAUDE.md, skills, commands, plan mode, permissions, and CI/CD.
- Engineer prompts for precision, few-shot consistency, JSON schemas, validation, retries, and structured output.
- Manage context windows, summaries, provenance, uncertainty, human review, and reliability tradeoffs.
- Apply architecture judgment to customer support agents, research pipelines, CI review, extraction, and developer tools.
- Practice CCAR-F scenario questions across all domains with quizzes and two full-length mock exams.
Prerequisites
- Python programming experience is recommended.
- Command-line experience is expected, including running tools, editing files, and working in developer environments.
- Some experience building with LLM APIs, Claude Code, Agent SDK, or MCP is helpful, but the course teaches from first principles.
- This is a technical architect-level course and is not intended as a beginner introduction to Claude or generative AI.
Target Learners
- Solution architects and senior engineers preparing for the Anthropic Claude Certified Architect – Foundations exam.
- Developers and technical leads building production applications with Claude, Claude Code, Agent SDK, Claude API, and MCP.
- AI engineers designing agentic systems, multi-agent workflows, tool integrations, structured output pipelines, or Claude-powered CI/CD.
- Technical learners who want runnable labs, exam-style practice, and production-oriented tradeoff guidance rather than a high-level overview.
Optimize your Udemy Learning Environment 0:54
Intro: The Exam 1:05
The Exam: What Architects Decide 6:11
The Exam: Deterministic vs. Probabilistic Controls 4:12
The Exam: Domains and Scenarios 4:52
Intro: The Claude Messages API 1:07
The Claude Messages API: Stateless Requests & Messages Array 7:19
The Claude Messages API: Content Blocks & stop_reason 4:36
The Claude Messages API: Request Controls & Output Limits 6:25
Lab: Prerequisites 8:12
Getting Course Material + Repository 1:00
Intro: The Tool-Use Loop 1:04
Tool-Use Loop: Why Tools Exist 5:16
Tool-Use Loop: The Tool Definition 3:13
Tool-Use Loop: Complete Cycle 7:33
Tool-Use Loop: : Choosing The Next Tool Call 8:13
Intro: Tools That Get Selected 1:11
Tool Selection: Description 4:53
Tool Selection: Diagnosing Misrouting 6:14
Tool Selection: Splitting & Scoping Tools 7:00
Tool Selection: Built-In Tools Grep, Glob, Read, Write, Bash 6:17
Intro: Model Context Protocol (MCP) 1:13
MCP: What MCP Solves 7:40
MCP: Tools, Prompts and Resources 4:00
MCP: Building A Server 5:17
MCP: Structured Error Responses 4:46
MCP: Integrating Servers 5:08
Lab 1a: Build MCP Server and Connect to Claude Code 12:24
Tool Design and MCP Integration (Exam domain 2)
Intro: The Agent SDK 1:04
Agent SDK: The Loop as a Library 9:08
Agent SDK: Built-In & Custom Tools 4:53
Agent SDK: Hooks & Subagents 5:41
Agent SDK: Conversations & Sessions 5:28
Agent SDK: Loading From Disk 3:57
Agent SDK: Permission Mode & Allowed Tools 7:10
Agent SDK: Scoped Tools & Cross-Role Exception 4:12
Lab 1b: Build Agent + Hooks with Agent SDK + MCP Integration 14:49
Intro: Orchestration 1:13
Orchestration: Multi-Agents Coordinator & Subagents 5:27
Orchestration: Defining Subagents 3:44
Orchestration: Passing Findings Explicitly 4:36
Orchestration: Parallel Subagents 3:46
Orchestration: Subagent Routing 4:16
Orchestration: Recover Locally & Report What Remains 4:52
Lab 2: Multi-Agent Research Pipeline 10:50
Intro: Enforcement and Handoff 1:12
Enforcement & Handoff: Hooks 3:47
Enforcement & Handoff: Normalise Tool Results 3:48
Enforcement & Handoff: Gates 3:52
Enforcement & Handoff: Structured Handoff 3:18
Enforcement & Handoff: Verify & Investigate in Parallel 3:25
Lab 3a: PreToolUse and PostToolUse Hooks for Support Agent 10:13
Inro: Decomposition and Sessions 1:08
Decomposition & Sessions: Fixed Chain or Adaptive Plan 8:06
Decomposition & Sessions: Resume, Fork or Restart 4:22
Lab 3b: Integrate Hooks into Support Agent with Agent SDK 8:39
Agentic Architecture and Orchestration (Exam domain 1)
Intro: Configuration That Gets Followed 1:03
Configuration: CLAUDE.md 6:24
Configuration: Writing CLAUDE.md Rules 3:37
Configuration: Splitting CLAUDE.md 3:48
Configuration: Path-Specific Rules 4:58
Intro: Skills and Commands 1:10
Skills & Commands: Skills as Procedures 5:49
Skills & Commands: Slash Commands 3:00
Skills & Commands: CLAUDE.md vs. Skill vs. Hook 5:22
Intro: Working Modes 1:11
Working Modes: Direct Execution vs. Plan Mode 4:57
Working Modes: Permission Modes 3:52
Working Modes: Iterative Refinement 5:43
Intro: Automation 1:09
Automation: Headless Mode 5:52
Automation: GitHub Action 4:08
Automation: Repeating Reviews 4:43
Automation: Testing Standards 4:13
Lab 4: Claude Code Headless as GitHub Action 9:20
Claude Code Configuration and Workflows (Exam domain 3)
Intro: Precision 1:10
Precision: Explicit Criteria 4:36
Precision: Precision vs. Coverage 3:44
Precision: Few-Shot Examples 6:10
Precision: Document Extraction 4:39
Intro: Structured Output 1:10
Structured Output: Documents to JSON 3:52
Structured Output: Extraction as a Tool 5:22
Structured Output: Missing Fields 2:53
Structured Output: Schema Validation 2:54
Structured Output: Grounded Confidence Score 3:06
Structured Output: Learning From Dismissed Findings 2:59
Lab 5: Document Extraction Pipeline with Messages API 11:06
Intro: Scale and Review Architecture 1:14
Scale & Review Architecture: Message Batch API 3:27
Scale & Review Architecture: Batch vs. Synchronous 2:16
Scale & Review Architecture: Multi-Pass Review 3:37
Scale & Review Architecture: Multi-Instance Review 3:16
Scale & Review Architecture: Test Small & Batch Corpus 5:07
Lab 6: Claude Code Direct Execution, Plan Mode & Fork/Resume 13:05
Prompted Engineering and Structured Output (Exam domain 4)
Intro: Context as a Budget 1:10
Context Management: Context Limit 6:07
Context Management: The Levers 5:57
Context Management: Directed Compaction 4:06
Context Management: Verbose Discovery & Context Degradation 4:34
Context Management: Recovering State 2:46
Intro: Deciding Under Uncertainty 1:20
Uncertainty: Escalation Triggers 3:39
Uncertainty: Ask vs. Escalate 2:37
Uncertainty: Human Review & Calibration 3:08
Uncertainty: Provenance & Conflicting Sources 3:10
Uncertainty: Meaningful Research Reports 3:41
Context Management and Reliability (Exam domain 5)
Intro: Exam Readiness 1:08
Exam: Sample Question 3:48
Exam: Domains Recap 11:42
Exam: Registration 1:46
Practice Exam 1: Full-Length Exam Simulation, CCAR-F
Practice Exam 2: Full-Length Exam Simulation, CCAR-F
Bonus Lecture 1:51
Access Files No pi required.
Free Course files 1 file package
- Full Pack
Discussions are closed
Comments are currently disabled for this course.
![AI Glossary at Work: A Practical Guide for Everyone [2026]](https://coursepi.com/wp-content/uploads/2026/10/ai-glossary-at-work-a-practical-guide-for-everyone-2026-300x169.webp)
