Technical articles on Claude, agentic AI, and certification preparation.
A practical guide to building AI agents with Claude. Covers the agentic loop lifecycle, single-agent vs multi-agent patterns, the coordinator/specialist architecture, the Claude Agent SDK, and what Domain 1 of the certification exam actually tests.
A practical guide to the Model Context Protocol. Explains how MCP servers, clients, and transports work — with runnable code examples and architectural patterns for real production use.
Technical walkthrough of CLAUDE.md configuration, the file hierarchy, modular @import organisation, and how Claude Code uses these files during sessions.
Guide covering plugin commands, structure, bundling, and marketplace integration. Explains how plugins extend Claude Code with domain-specific tools and slash commands.
General availability announcement with implementation examples, the new API parameter format, and what structured outputs mean for the exam's prompt engineering domain.
Hands-on techniques including system prompt design, few-shot examples, output validation patterns, and how to avoid common anti-patterns like prompt injection.
Strategies for preserving data integrity in long sessions without progressive summarisation losses. Covers rolling summaries, anchoring, and RAG as alternatives.
Comprehensive exam preparation covering all five domains with study strategy recommendations, time allocation advice, and the most commonly tested concepts.
Overview of new features relevant to architects: expanded context windows, improved tool use reliability, and updates to the Agent SDK.
A balanced, technical comparison of Claude and ChatGPT for developers. Covers API design, context windows, tool use, structured output, coding performance, and pricing.