mattyhorne.co.uk
Whiteout issue 01

Building useful systems with AI, in public.

IT Manager at LF Europe. Founder of Brumalia Studios. Building useful systems with AI in public. I work where infrastructure, AI tooling, and product thinking overlap — documenting what gets built, what breaks under pressure, and what turns into something worth keeping.

Current lanes
LF Europe — infrastructure, AI tooling, ESG, and Digital Product Passport delivery.Brumalia Studios — websites, apps, automations, and agent systems for real work.Writing — what survives contact with reality gets documented here.
002
tutorialai-agents27 February 20268 min read

We built Mission Control using a multi-agent architecture. Here's the exact system we use: workspace structure, pipeline, memory, and what actually works in production.

004
reflectionai-agents7 May 20267 min read

This week wasn’t about one breakthrough feature. It was about a chain of quieter realisations: that workflows need proof, not just orchestration; that internal patterns need packaging, not just admiration; and that a stalled pipeline can teach you as much as a working one.

Four entries that explain the arc: why this site exists, how the agent stack works, what Mission Control became, and where the whole thing starts looking like a business.

Browse all writing →

These are not retrospective case studies. They are live dossiers: projects being scoped, shaped, and pressure-tested in the open.

01
Discoverylead-genclient-work

Polyurethane Lead Gen App

Pre-spec

A lead-generation and qualification system for a polyurethane products business — research first, then a disciplined MVP spec.

Current focus: Business discovery, ideal customer definition, sales-process mapping, and agreement on what success should look like before building anything.

02
Use-case definitionfootballanalytics

HTAFC Women’s Analysis Tool

Planning

A football analysis and research tool for Huddersfield Town AFC Women, beginning as a focused insight product with commercial potential later.

Current focus: User definition, data-source discovery, and deciding whether v1 should be reports, opposition analysis, player insight, or an even tighter wedge.

03
Product framingfinanceai

Personal Finance AI App

Planning

An AI-first personal finance app aimed at making money clearer for normal people, not just spreadsheet obsessives.

Current focus: Feature definition, target-user clarity, market-gap research, and a firm v1 boundary that stops the product trying to do everything.

Open the project index →
01Turning Brumalia Studios from a promising experiment into a studio with clear offers, sharper positioning, and real client delivery.
02Building and documenting AI-agent workflows with Hermes, Claude Code, GBrain, and n8n — with more proof than hype.
03Delivering ESG and Digital Product Passport work across LF Europe, Orrsum, and Blackcat Fireworks without losing the operational plot.
04Scoping three live product threads properly: lead generation, women’s football analysis, and personal finance.

Lab

Rougher operating views, experiments, dashboards, and internal tools that belong on the site because they prove the work is real.

Featured viewLLM cost board — live lane discipline, spend visibility, and routing policy.Open the lab →

If you want sharper systems, cleaner delivery, or a practical AI build, start here.

I’m interested in work that needs judgment, not just output: websites with a point of view, internal tools that reduce drag, and automations that save real time.

Discuss a buildView GitHubConnect on LinkedIn