ai-first software builder — london, uk
I build AI agents, automation, websites and apps, plus Web3 systems that move real money. Python, TypeScript, Rust, C#: whichever the job needs.
01services.py
Most briefs land in one of these. The good ones land in two at once: an agent inside a product, or a payment flow with a chain behind it.
def agents():01
I wire AI agents into the tools a business already runs (inbox, CRM, docs, payments) with scoped access, a log of what they did, and a human approval step where it matters.
def automation():02
The copy-paste between tools becomes a pipeline that runs on its own and tells you when something needs a person.
def products():03
Websites, e-commerce, dashboards, portals and mobile apps, built as one piece rather than stitched together.
def bridge():04
Solana programs in Rust and Anchor, plus the app, wallet flow and dashboards around them, so the chain does its job and stays out of the way.
02sectors.py
The sectors I’ve built for, and what usually gets built in each.
03bridge.py
The interesting Web3 work connects to things outside the chain: a fundraiser, a creator, a customer who has never held a wallet. I build the programs, the verification, and the normal-looking product on top.
A token launchpad where trading fees go to an attached GoFundMe campaign: real-world fundraising carried by on-chain activity.
A launchpad that sends a token’s trading fees to the creator attached to it, so the people building attention share in it.
Simplified example: a trade fee is split, the cause's share is credited to escrow, and payouts are scheduled once the campaign is verified.
04stack.py
What each language is for, then the tools by layer. Pick a layer to print it.
05process.py
One person from the first call to production, so nothing gets lost in a handover.
A short call: the problem, what already exists, and what done looks like.
What I’ll build, and just as clearly what I won’t.
How it looks, where the data lives, and what it connects to.
Working software you can click as it comes together.
Deploy, watch it run with real use, and fix what shows up.
AI handles first drafts, boilerplate and tests. I make the decisions, review the result, and say so when something isn't worth building.
06work.py
Ten builds on one wheel: a fraction of what I’ve shipped. Credits are exact; solo means solo, co-dev means shared.
Live Solana token intelligence: on-chain data, AI analysis, bundle detection, holder intel, wallet tracing, and risk scoring in one pass.
07practice.py
Business software is the main event. On-chain work and hand-made craft are the same instinct pointed at different material.
on-chain/
The Solana side of the work is public by nature. A few places it has turned up, and who I build with.
Featured on Solana Floor.
# solana ecosystem coverage
Shipping with a Solana gaming dev team.
# not solo — building games with the people building them
Inside Seeker’s Summer Quests.
# solana mobile campaign
craft/
3D, image and brand work, two pixel-art collections, and pieces that end up stitched under neurothreadz.
3D in Blender, image and brand work in Photoshop and the Adobe suite, and generative pieces built in code.
Custom stitched apparel, patches and one-off pieces. My own label, and the physical end of the practice.
# nft collections
Want something drawn, stitched, or minted rather than deployed? leave a signal
08now.py
A quick read on what’s being made, maintained, and listened to right now.
# whoami
I’m an engineer and maker in London. This page is the public part: applications, protocols, games, visual systems and stitched objects.
I like owning a thing from its first messy note to the part where it actually runs: interface, API, database, payments, deployment and the machine underneath.
Outside client work it’s privacy-first infrastructure, on-chain tooling and small ideas worth making physical. neurothreadz is that last part: artwork translated into thread.
09faq.py
Something not covered? Send a message and I’ll give you a straight answer.
Send a message through the form below. If it looks like a fit, I’ll suggest a short call. If I’m not the right person for it, I’ll tell you that instead.
It depends on scope, so I won’t guess here. Once I understand the problem, you get an honest estimate for both.
Yes. Most of what I build for companies and agencies is private, which is why no client is named on this page. I’m happy to sign one before going into detail.
It depends on the provider and its settings, so I choose both with your data in mind and keep secrets out of prompts. If data can’t leave your environment, local models are an option.
Usually not. I only suggest on-chain pieces when they solve something a normal database can’t.
10contact.py
Fill in the call and run it. It lands straight in my inbox.