I run three things in parallel, and they all feed this course.
AIFirstPartner is my consulting practice. I've shipped 12 production AI systems through it for clients in government, legal, and enterprise. Most of the patterns in this course come directly from that work: what actually holds up when the prompt meets real users, real data, and a real on-call roster.
I'm also fractional CTO at Speculo, a cyber risk software company, where I'm responsible for the technical direction of the product. Building AI features into a security product teaches you very quickly which shortcuts are safe and which ones will come back to bite you.
My day role is managing cloud and data platform engineering for a large Australian government department supporting 20+ agencies across the portfolio. 15+ years of enterprise technology sits behind that, and it's the reason I write about evals, reliability, and cost as if they matter, because at that scale they do.
I publish my build process openly at 150days.aifirstpartner.com. If you want to see how I work before you buy anything from me, that's the place to read.
Alongside the consulting work, I build a few of my own tools. You might know them:
The session analyser for AI-assisted development. Upload a Claude Code transcript and get scored across 8 fluency dimensions, with actionable recommendations.
A VSCode 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 hearing the same question from engineers: "I can call the Claude API, but how do I actually build something real with it?" This course is my answer.
Building with the Claude API is a practical, code-first course for software engineers who want to integrate Claude into real applications. It covers everything from your first authenticated request to building production tool systems, hybrid RAG pipelines, MCP servers, and agentic workflows.
Every concept ships with working Python code you can run and modify. You'll build a multi-turn chat system, a prompt evaluation pipeline, a complete tool use system, a hybrid RAG pipeline, and an MCP server and client over the course of 13 modules.
Most LLM tutorials stop at "here's how to call the API." That's where the real work begins. Production applications need evaluation, error handling, structured output, tool use, retrieval, and a clear story for when to use a workflow vs a full agent.
I built this course to fill that gap. It takes engineers from "I made an API call" to "I shipped a Claude-powered feature into production" without skipping the parts that matter: evals, tool design, context management, and pattern selection.
Everything is grounded in the official Anthropic API documentation and patterns we use in our own production systems.
Questions, feedback, or just want to say hi — I'd love to hear from you.
All tools built by AI First Partner