Learning Hub
5 domains · 30 lessons covering everything you need to pass the Claude Certified Architect (Foundations) exam.
Your progress
Start learning to track your progress across all 30 lessons.
Diagnostic Assessment
Not sure where to start? Take a 15-minute diagnostic to find your strengths and gaps across all 5 domains.
Agentic Architecture & Orchestration
Design, build, and reason about multi-agent systems. Understand agentic patterns, the Claude Agent SDK, and orchestration strategies.
7 lessons
Introduction to Agentic Systems
What makes a system 'agentic' and when to use agents vs. simple prompts.
Prompt Chaining Pattern
Sequential multi-step workflows where each step feeds the next.
Routing & Parallelization Patterns
Classify inputs to specialized handlers and run tasks concurrently.
Orchestrator-Workers Pattern
Central orchestrator delegating to specialized sub-agents.
Evaluator-Optimizer Pattern
Iterative loops with evaluation and refinement for quality output.
Claude Agent SDK Deep Dive
Building custom agents with the Claude Agent SDK — the agent loop, tools, and hooks.
Multi-Agent Communication & Handoffs
How agents pass context, results, and control between each other.
Tool Design & MCP Integration
Design effective tools for Claude, master the Model Context Protocol, and build MCP servers.
6 lessons
Tool Use API Fundamentals
Defining tools, tool choice modes, and multi-tool orchestration.
Tool Design Best Practices
Writing effective tool descriptions, scoping, and error handling.
MCP Architecture & Primitives
Understanding MCP hosts, clients, servers, and the three primitives.
Building MCP Servers
Creating MCP servers with FastMCP (Python) and TypeScript SDK.
MCP Security & Production Deployment
Authentication, rate limiting, and production strategies for MCP.
Tool Boundary Management
Managing tool complexity and preventing model overload.
Claude Code Configuration & Workflows
Master Claude Code as a development and automation tool — CLAUDE.md, slash commands, CI/CD, and team workflows.
6 lessons
Claude Code Overview & Core Capabilities
What Claude Code is and how it works as an agentic coding tool.
CLAUDE.md Hierarchy & Configuration
Project, directory, and user-level CLAUDE.md files and resolution order.
Custom Slash Commands & Agent Skills
Creating reusable commands and skill definitions.
Settings & Permission Management
Configuring settings.json, allowed tools, and MCP servers.
Claude Code in CI/CD Pipelines
Non-interactive mode, tool restrictions, and pipeline integration.
Git Integration & Team Workflows
Claude Code's git safety, plan mode, and team configuration.
Prompt Engineering & Structured Output
Master prompting techniques for reliable, production-quality outputs from Claude.
6 lessons
Core Prompting Techniques
Clarity, role prompting, few-shot, chain-of-thought, and XML structure.
Structured Output Enforcement
JSON output, prefilling, schema validation, and retry loops.
Data Extraction Patterns
Extracting structured data from unstructured text reliably.
Extended Thinking & Complex Reasoning
Using thinking blocks and budget tokens for complex tasks.
Output Determinism & Consistency
Strategies for consistent outputs — temperature, examples, constraints.
Prompt Testing & Evaluation
Evaluating prompt quality, A/B testing, and building eval frameworks.
Context Management & Reliability
Manage Claude's context window effectively and build reliable, production-grade systems.
5 lessons
Context Window Management
Token limits, prioritization, summarization, and sliding windows.
Long-Document Processing
Chunking, map-reduce, and hierarchical summarization patterns.
Prompt Caching Strategies
Reducing latency and cost with prompt caching.
Guardrails & Safety Mechanisms
Input/output validation, content filtering, and hallucination detection.
Error Recovery & Graceful Degradation
Handling API errors, timeouts, rate limits, and unexpected outputs.
Ready to test your knowledge?
Work through the lessons above, then put yourself to the test with the mock exam.
Take Mock Exam