I approach growth like a developer: understand the system, find the constraint, build the test, inspect the result, and iterate + scale.
Twenty years across agencies, corporate teams, startups, gambling, finance, and crypto. I work across the commercial and technical sides of a problem: customer psychology, acquisition economics, campaigns, data, automation, and code.
I find growth opportunities. Then I build the test.
Website · Projects · Writing · LinkedIn · X
My background includes black-hat marketing and adversarial growth campaigns for gambling operators. That experience shaped how I evaluate incentives, competition, measurement, platform dependencies, and the durability of an apparent advantage.
It also shaped how I think about risk management. I evaluate the opportunity alongside the exposure: what the approach depends on, how it could fail, what we stand to lose, and whether the upside justifies it.
Before scaling something, I want to know:
- Is the growth real? Customer quality, retention and commercial value matter alongside volume. Attribution, incentives and abuse can distort the numbers.
- What could break? Identify critical dependencies and the budget, operational and reputational exposure if they fail.
- When do we stop? Define the conditions for changing course before enthusiasm becomes the operating model.
A campaign can look brilliant in a dashboard and still be a terrible business decision.
| Area | What I work on |
|---|---|
| Performance & growth | Paid acquisition, SEO/GEO, creative testing, conversion, distribution, partnerships and retention. Connecting channel performance to the business underneath it. |
| Commercial diagnosis | Customer behaviour, offers, funnels, acquisition economics and measurement. Finding which assumption deserves a test. |
| Technical implementation | Tools, API integrations, data workflows, reporting, publishing systems and automation. Building the missing piece and connecting it to the work people actually do. |
| Applied AI | AI-assisted software development, agent workflows, research and information processing. Turning a useful capability into something people can use. |
| Risk & resilience | Questioning dependencies, failure modes, incentives and downside before committing more budget or operational trust. |
I've been automating work for years, long before it came with an AI label.
A desktop media manager for turning saved videos and audio into transcripts, summaries, and a searchable library. Local processing options or cloud AI providers. Saving something is easy. Finding the useful part a few months later is the problem.
A browser extension for detecting and inspecting hidden Unicode characters. It reveals text that ordinary rendering conceals, making the underlying content easier to examine.
An AI-assisted software directory that became an investigation into search visibility and AI citations. There's a... "follow-up at scale" in the making ;]
Make the problem inspectable - Define the behaviour we want to change, the constraints, and the evidence we need. “Improve growth” is a direction. It still needs a testable question. Build the smallest useful intervention - A campaign, landing page, integration, automation or small product. Choose the form that lets us learn something worth knowing. Make the result observable - Check the tracking, inspect the data and distinguish what happened from what we'd like to claim happened. Keep enough context to explain (and replicate) a result later. Work in short feedback loops - Put something in front of reality, inspect the response and improve the next attempt. Record the decisions and leave the useful work understandable to whoever touches it next. Know the failure cost - Small experiments should have bounded exposure. As something becomes important to the business, reliability, dependencies and recovery deserve more attention.
I like ambitious ideas. I also like logs.
I started with BASIC on a Commodore 64 at around ten. The recurring question was: how does this work, and what else can I make it do? That curiosity carried into twenty years of marketing, automation and commercial experimentation, across agency, in-house and startup environments. I've worked remotely and across countries for much of that time.
Currently, I work in performance marketing and contribute to Agentic Experience (experimental frontier of "marketing for AI agents").
Bring me a growth problem, a promising idea or a result you don't trust. I'm interested in work that needs commercial judgement, technical investigation and someone willing to build what the test requires.