I founded AIFirstPartner, a consulting practice with clients in government, legal, and enterprise. I've shipped 12 production AI systems to date. Individual courses like this one are not Anthropic products and have not been reviewed or endorsed by them.
Alongside the consulting practice, I'm fractional CTO at Speculo, a cyber risk software company. My day role is managing cloud and data platform engineering for a large Australian government department supporting 20+ agencies across the portfolio. That's 15+ years in enterprise technology behind the AI work.
I publish my build process in public at 150days.aifirstpartner.com. If you want to see how I actually work with Claude Code day to day, that's the place.
You might also know two of the tools I've built for the Claude Code community:
The session analyser for AI-assisted development. Upload a Claude Code transcript and get scored across 8 fluency dimensions, with actionable recommendations.
A VS Code extension that monitors your Claude Code sessions in real time: live health scores, anti-pattern alerts, and per-model cost tracking, right inside your editor.
Building these tools, I kept noticing the same pattern: the developers who get the most leverage from Claude Code aren't the ones with the cleverest prompts. They're the ones who know when to delegate work into a subagent and when to keep it in the main thread.
Introduction to Subagents is a focused micro-course on one of the most underused features in Claude Code: subagents — specialized assistants that run in their own isolated context windows, do focused work, and return just a summary to your main thread.
In four short lessons you'll learn what subagents are, how to create one with the /agents command, four design patterns that make them effective, and a decision framework for when to use them versus keeping work in your main conversation.
The whole course takes about 30 minutes and is completely free.
Most developers I talk to either don't know subagents exist, or they've tried them once and given up because the results felt unpredictable. The truth is that subagents are powerful, but they have a learning curve — you need to understand context isolation, structured output formats, and which tasks actually benefit from delegation.
I built this course to flatten that learning curve. Every lesson is grounded in real Claude Code workflows: a code-quality reviewer subagent runs as the worked example throughout, and the final lesson distils the use-for / avoid-for tradeoffs into a single decision rule you can apply on the fly.
If you use Claude Code regularly, this is a 30-minute investment that should pay back the first time you delegate a research task and keep your main context clean.
Questions, feedback, or just want to say hi — I'd love to hear from you.
All tools built by AI First Partner