About
I’m Mark. My background spans streamed data, machine learning, web applications and enterprise application security. Most of my side projects grow out of the overlap between those.
What I build
Most of my projects start with a question, and the answer turns out to need some data that nobody has collected yet:
- Is the nation feeling festive yet? That became How festive is the UK?, a daily festivity index built from Google searches for All I Want for Christmas Is You.
- How do you keep an AI agent up to date with what’s happening in the world? That became Follo, which watches news sources for the topics you care about and pushes the relevant events straight into your workflows.
- Who holds data about us that we can’t change? That became haveibeencloned, a guide to biometric and genetic data risk with a news tracker going back to 2019.
- Can you spot a phishing link from the address alone? That became RiskyClick, a link checker that runs in your browser, trained on the phishing links I collected from 2020 to 2025.
- Do Bargain Hunt teams actually make money? That became Bargain Hunt in numbers, which uses an LLM to pull prices out of subtitles. (Mostly yes. Jewellery is the safe bet.)
The same things tend to come up each time: getting messy sources into a shape you can trust, checking the results rather than taking them on faith, and keeping security and privacy in mind from the start.