Solutions Architect · UK · since 2020

Connected data, AI systems & IT, built end-to-end.

25 years in IT, infrastructure and systems; five years deep in AI. I build what doesn’t exist yet, fix what’s quietly breaking, and rebuild the legacy software businesses still depend on, modern, connected and better.

Start a project →Read the approach

Hover to trace · drag a cluster to move it

What I build

A stack of solutions, built end to end.

One person, one connected system, the whole build. From the AI that does the work to the platform it runs on. This isn’t advice, it’s action.

Start a project →

AI systems & automation

Agents and pipelines that run themselves, and the infrastructure that lets any AI plug into a business and get real work done.

Full automation, end to end
The universal system AI runs in

See the approach →

BUILD

Platforms, SaaS & web

Secure, connected and fast. Multi-org platforms, web apps and sites, plus legacy software rebuilt modern and better.

Multi-org SaaS & platforms
Legacy rebuilt, faster

Start a build →

NATIVE

Software & private AI

Native Windows and macOS software, browser extensions, and private on-device AI for work that cannot leave the building.

Mac & Windows apps + extensions
On-device, private AI

Talk it through →

DATA

Connected data & compliance

Immersive 3D knowledge graphs that map your connected data, and legal, compliance-grade systems that hold up under scrutiny.

See your data as a graph
Solicitor-grade confidentiality

Explore the graph →

Selected work

Proof, not promises.

All case studies →

AI Innovation

Claude Running Claude: A Real-Time AI Agent Network, Before Native Multi-Agent

A working real-time communication network between AI agents, Claude orchestrating other Claudes to coordinate on the same project in real time, with shared context. Built and running before Claude had native multi-agent support.

View case study →

Web Development

From a Flat Page to a Full E-Commerce Platform: One System, Built Fast and Audited Hard

One system that builds the whole range, fast flat sites, content-managed sites, and complete e-commerce, on Next.js, Vercel, Supabase and AWS. Spec-driven, OWASP-audited and documented, so every site is fast, secure and genuinely maintainable.

View case study →

Enterprise Development

Running 24 Concurrent Claude Code Instances

One AI in a terminal is a tool. Twenty-four running at once, concurrently, under one controller is an operation, and I was running them that way before the tooling was meant to make it possible.

View case study →

Business Growth

Beast Mode: 5000% Growth in Twelve Days

A focused twelve-day autonomous build sprint, and, at the time, a 5000% jump. What it actually takes to let AI run for hours instead of babysitting it line by line.

View case study →

SEO Strategy

The SEO Archive That Tripled Long-Tail Traffic

Most content gets published once and buried. An archive built so older work keeps earning its place in search, and tripled long-tail traffic by doing it.

View case study →

Innovation Methodology

Five-Day Build Cycles for What the Industry Quotes at Three Months

The work a traditional team pencils in for three months, shipped in five-day cycles with something real to show each day. Not a shortcut, an AI-first cadence the old timeline was never built for.

View case study →

Performance Optimisation

64 Images, 70% Lighter, Zero Quality Lost

Sixty-four images, re-engineered for a 70% drop in weight with no visible loss of quality. The easiest performance win on most sites, taken properly.

View case study →

SEO Automation

The SEO Optimiser MCP: 15 Hours a Month, Automated Away

Good SEO is mostly the same checks, run again and again. A tool that collapses that repetitive work into a single command, and gave back about 15 hours every month.

View case study →

Content Publishing

Blog Publisher v3: Markdown Fixed at the Source, Across 28 Posts

Broken formatting in a published article is a small bug with an outsized cost to credibility. The third version of my publishing tool fixed it at the source, and cleared it across 28 posts.

View case study →

Whitepaper

AI First Principles: Structure as Substrate for Drift-Free, Model-Agnostic Software Construction

Treat AI-built work as a file-and-folder program: put the guarantees in the structure, not the model’s memory; assign each task to the cheapest correct runtime; AI first, human second, which makes the AI an interchangeable runtime.

Read the whitepaper →