About

About Alex Dunlop

I'm a Kiwi developer and writer based in Sydney. I've spent the last 10 years building AI products, and I write about what I learn along the way.

I got into AI before the current wave. Before ChatGPT, before agents, before “vibe coding” was a phrase. I was building natural language generation systems at Arria NLG and figuring out how to make machines write useful English. That background shapes everything I do now.

Today I work across startups, consulting, and open source. I write about AI coding, developer workflows, and what it actually looks like to ship software with agents. 1.8 million views on Medium and growing. Not because I'm good at SEO. Because engineers keep coming back for the practical stuff.

Alex Dunlop smiling and giving two thumbs up in an office workspace while wearing a black jacket, hoodie, and backpack.

What I do

I build AI products. Full stack. Frontend, backend, AI layer, infrastructure. Not the “I fine-tuned a model once” kind. The “millions of real users in production” kind.

I also write. A lot. Deep dives on Claude Code, Cursor, AI agents, and the workflows that actually work when you're trying to ship. My writing has reached over 1.8 million views because I focus on what's useful, not what's trendy.

And I'm building a course. 10 years of AI product engineering distilled into a structured path for engineers who want to get serious about shipping with AI. Join the waitlist here.

Right now

I'm a Senior Engineer at Popp AI, building the conversation engine that powers recruitment for some of the largest enterprises in the world. Think millions of candidate interactions, real AI at production scale.

I also run Cub Digital, the consultancy I co-founded in New Zealand. We've built AI solutions for the All Blacks, the NZ Olympics, Air New Zealand, and the NZ Stock Exchange. Different problems, same approach. Make it useful. Make it clear. Ship it.

Outside of work, I volunteer as a Senior Product Engineer for Aruuri, a community fund run by Somalis in Tech that pools diaspora contributions to back vetted projects across the Horn of Africa. Healthcare, education, livelihoods. Everything is tracked through a public ledger. Every project goes through proper due diligence before a dollar moves. It's one of the most meaningful things I get to put my skills toward.

When I'm not shipping, I'm writing the newsletter (Dunlop's Dev Log), working on a course for engineers who want to level up with AI, and occasionally making house music remixes that absolutely nobody asked for.

Ventures

A few places where the work shows up.

Popp AI

Conversation engine for enterprise recruitment. Powers AI-driven candidate interactions for companies like Brunel and Robert Walters at massive scale.

Cub Digital

AI consultancy I co-founded. Clients include the All Blacks, Air New Zealand, NZ Olympics, and the NZ Stock Exchange.

Aruuri

Community fund pooling diaspora contributions to back healthcare, education, and livelihood projects across the Horn of Africa. Full public ledger. Structured due diligence. I volunteer as a Senior Product Engineer.

The path here

  1. Arria NLG

    Where it started. I built natural language generation systems before LLMs existed. We were making AI write reports, summaries, and analysis from raw data. This is where I learned how language models think, years before GPT made it mainstream.

  2. Cub Digital

    Co-founded an AI consultancy in New Zealand. Built solutions for the All Blacks, Air NZ, NZ Olympics, and the NZ Stock Exchange. Learned how to take AI from “cool demo” to “thing that runs in production and doesn’t break.”

  3. Popp AI

    Joined as a senior engineer to build the conversation engine. Scaled from early stage through serious growth. The platform now handles AI conversations for some of the biggest recruitment companies in the world.

  4. Writing

    Started sharing what I was learning on Medium. Crossed 1.8 million views. Launched Dunlop’s Dev Log on Substack. The articles on Claude Code, AI agents, and developer tooling keep resonating because they come from actually doing the work, not summarising someone else’s thread.

  5. Now

    Building the course, writing the newsletter, shipping at Popp, volunteering at Aruuri, and figuring out how to help more engineers get to where I am with AI without spending a decade doing it.