Skip to content

    Certification

    CCA Developer — Course Overview

    Overview of the Claude Certified Developer — Foundations course: five modules from prompting and tools to Claude Code, MCP, evals and security.

    By Published 3 min read

    Share

    What This Course Is

    The Claude Certified Developer — Foundations course prepares you to design, build, and operate production Claude integrations. It covers the full developer stack: how Claude processes tokens and context, how to structure prompts and tools reliably, how to use Claude Code and MCP servers in real workflows, how to run evals and harden against security threats, and how to choose and deploy Claude across platforms.

    This is a practitioner-level course for developers who are building with Claude, not just using it. Every module connects theory to code.


    Who This Is For

    • Developers building Claude integrations for production
    • Engineers preparing for the Claude Certified Developer certification
    • Technical leads evaluating Claude for enterprise adoption
    • Anyone who has shipped something with Claude and wants to close the conceptual gaps

    Prerequisites: You should be comfortable reading Python or TypeScript and have made at least one Claude API call. This course is not an introduction to prompting — it starts one level deeper.


    Five Modules

    Module Topic What You'll Learn
    M1 MSO Foundations Tokens, context window, sampling, model family, reasoning, SDK vs REST
    M2 Production Prompting, Agents & Tool Use System prompts, XML, few-shot, extended thinking, tool-use loop, agent patterns, memory, streaming
    M3 Claude Code, MCP & Integration Permission modes, CLAUDE.md, hooks, subagents, MCP servers, enterprise auth
    M4 Production Engineering, Evals & Security Evals, testing, failure handling, model selection, cost, prompt injection, ZDR
    M5 Accelerators, IP & Deployment Packaging accelerators, contributing back, platform comparison, model versioning, trust boundaries

    Plus: a practice exam (53 questions) with a second practice set for a fresh run, domain-by-domain practice quizzes so you can drill your weakest area specifically, and a rapid-reference cheatsheet.


    How to Use This Course

    Read in order. Each module builds on the previous. M2 assumes you know M1's context-window mechanics. M3's hook examples assume you've read M2's tool-use loop.

    Follow the code. Every code block is production-ready. Don't skim them — the exam tests whether you understand the implementation details, not just the concepts.

    Use the callouts. [!NOTE] callouts flag exam-critical distinctions. [!TIP] callouts give practical rules of thumb. [!WARNING] callouts flag traps that produce plausible but wrong answers.

    Take the practice exam last. Work through all five modules, then take the 53-question quiz. For every wrong answer, find the relevant section in the module and re-read it.


    Certification Notes

    The Claude Certified Developer exam tests:

    • Conceptual accuracy — can you explain why a design decision is correct, not just that it is?
    • Diagnostic thinking — given a symptom (wrong output, tool not called, context overflow), which fix applies?
    • Boundary awareness — what is Claude's responsibility vs the application layer's?
    • Security judgment — which design is safe? Which one has a structural flaw?

    The exam does not test memorised facts. It tests whether you have built the right mental model.


    TIP

    This is a multi-session plan, not a single sitting. Reading all five modules takes a solid chunk of time on its own, and between the 53-question exam, the eight domain-specific practice quizzes, and the second full mock exam, there's several more hours of practice material — plan on spreading this across a week of evening sessions rather than one sitting.

    Domain 1

    Agents and Workflows

    Tool-use loop mechanics, stop_reason semantics, workflow-vs-agent design, orchestration, memory strategies.

    Domain 2

    Applications and Integration

    Accelerators, requirements capture, systems lifecycle, deployment platforms, versioning, trust boundaries.

    Domain 3

    Claude Code

    Permission modes, configuration hierarchy, CLAUDE.md, hooks, subagents, and skills vs plugins.

    Domain 4

    Eval, Testing, and Debugging

    Eval design principles, grading methods, LLM-as-judge calibration, testing levels, tracing.

    Domain 5

    Model Selection and Optimization

    Tokens, the shared context window, sampling and non-determinism, the Claude model family, reasoning modes.

    Domain 6

    Prompt and Context Engineering

    The four prompting techniques, system prompts, XML tags, few-shot, structured output, extended thinking.

    Domain 7

    Security and Safety

    Prompt injection and defense in depth, trust boundaries and least privilege, ZDR/HIPAA/compliance.

    Domain 8

    Tools and MCPs

    MCP primitives (tools, resources, prompts), transport options, scope, and enterprise integration patterns.

    Ask about this article

    Get answers grounded in this post. AI-generated — based on this article, and may be imperfect.

    Free: CCA Foundations cheat-sheet (PDF)

    The domains, the 3 universal rules, core concepts, and exam-day shortcuts — one page. Enter your email and it's yours, plus my weekly AI-architecture notes.

    No spam. Unsubscribe any time.

    Scaled AI Weekly

    Enjoyed this? Get more like it every Monday.

    Real architecture decisions, LLMOps patterns that survive production, and engineering leadership advice — from 12+ years of building at enterprise scale. Free. No spam. Unsubscribe anytime.

    Get new articles by email

    Free. We email a confirmation link first, and you can unsubscribe from any email.

    Join engineers building production AI systems

    Comments