Home • About • Projects • Blog • Resources • Contact

Real-world experiments where I’ve applied product thinking to AI workflows, agent systems, or LLM tools.

💡 Featured Projects

Launchling

Launchling was a tiny tool to help non-technical founders start with clarity and confidence.

It was a simple, AI-powered app that helps early-stage, non-technical founders test startup ideas with a personalised plan. This proved an excellent vehicle through which to explore what building with AI actually demands once you get past the demo: not just prompting a model into producing plausible-looking output, but constructing an evaluation pipeline rigorous enough to distinguish a genuinely good plan from a merely fluent one, wrestling with schema and data structure long before any of it reached a user, and learning firsthand where an LLM's judgement could be trusted and where it still needed a human standing over it. It hardened a conviction that has shaped everything I've done with AI in a product role since: that the real work with these tools lives in evaluation and judgement, not generation.

→ Technical breakdown

• Built With: React, Vercel, GPT-4o, Airtable
• Status: Live MVP (Jul 2025)
• Tags: AI Product · Solo Founders

🪛 Tools I use

Notion · Super.so · Tally · Framer · React · Airtable · Zapier · Vercel · Custom GPTs · OpenAI GPT-4o · Cursor · Claude Code

Want the quick version?

Check out my Resources page for short notes on tools I’ve tested — what they’re for, what they’re good at, and where I’ve used them.

Prefer a deeper dive?

Head to the Blog for longer posts on how I used them, why I chose them, and what I learned along the way.