Tech lead & product builder. London.
I build products, explore frontier AI, and turn the interesting bits into things you can use.
Download CV (PDF)
CharGen
AI creation and campaign management for tabletop RPGs. Art, video, audio and worldbuilding, alongside session recaps, connected campaign records and an assistant grounded in your campaign history.
Step into CharGen
LOAF
One button. A tower of sleepy cats.
A small experiment in
making something
you’ll want to play one more time.
ZOOKWORKS
Build a creature from motors, legs and wheels, and teach it to walk. Then race it, push a rival out of a sumo ring, or send a friend a ghost to beat. Free, in your browser.
Build your first one
I like the point where
engineering meets
“could we?”
I’m a tech lead and software engineer in London, working across distributed systems, developer platforms and agentic AI. I lead teams and stay close to the architecture and code.
Outside that, I build. Creative tools, tiny games, useful experiments. I’m interested in what frontier models can do when you give them real problems, thoughtful direction and a way to check the result.
Three AIs, one very human questionGPT, Astra and Opus answer the meaning of life in three complete one-minute films.
Three models, one UFOClaude Opus 5.5, GPT-6 Sol and GPT-6 Luna make the same tiny game.
One brief, four bridgesOpus, Astra, GPT-6.1 Sol and GPT-6 Sol build four playable goblin bridge games.
Four models, one tavern trayGPT-6 Astra, Claude Opus 5.5, Sol and Luna make four playable fantasy taverns.
RunbookOne small app brief, tried with six AI models.
Beyond the first promptA follow-up experiment with design guidance.
Localhost tab labelsA browser helper for too many projects.
A date is not a timestampA reproducible timezone experiment.
One Bug, PleaseA small game about shipping software.
Pin one public GitHub pull request to its exact commit range, then end the review with one closed material-defect verdict.
Print a scope receipt or launch a Conversation Lifeboat with One Small Fix. Keep the result in editable Markdown.
A terminal log counter for the cost of reconnecting.
Find the weight and repetition in instruction files.
Explore the surprises in Unicode comparison.
Source-backed launch copy with something to point to.
Examine a subscription and get its cost or work time per useful session.
Play Token Herd while your agent runs.
