AI & Automation Lead · Solutions Architect · UK / Remote

I build the production AI
systems businesses run on.

Operator-engineer. I run a small UK services group day-to-day and design, ship and own the LLM and automation systems it depends on — RAG knowledge platforms, decision co-pilots, and an integration backbone where nearly half of all activity is now automated. Self-taught by shipping things a real business depends on every day.

Start a conversation → See the systems
~0.05→0.85
RAG recall@5, after rebuild
≈ half
of CRM activity automated
2×+
online conversion uplift
~3 yrs
self-taught → running tech
01

Selected systems

Production systems in daily use at a ~30-person, multi-£m UK services group. Figures are deliberately banded; specifics available under NDA.

/01

Multi-step LLM knowledge platform

Architected & built — hands-on

An organisation-wide retrieval system that turns scattered business knowledge into grounded, cited answers. Postgres + pgvector with hybrid (vector + keyword) retrieval and reranking; auto-capture; an automated eval harness and a human-in-the-loop critic gating what gets trusted.

recall@5 ~0.05 → ~0.85citation-groundedeval-gated
RAGpgvectorhybrid retrievalrerankingevalsPython
/02

AI decision-support suite

Designed, built & shipped

Task-specific co-pilots in daily use by leadership: a seasonal cashflow forecaster, an operations risk co-pilot that flags job risk from combined operational, weather and routing data, and marketing-attribution dashboards. Built on pure, unit-tested TypeScript libraries.

in daily use40+ unit testsdecisions on data, not gut
LLM co-pilotsforecastingTypeScriptdata viz
/03

Instant-quote & self-service booking

Architected the engine; led delivery

A central pricing engine as a single source of truth, powering AI-assisted instant quoting and a self-service booking journey that integrates CRM and e-signature. Moved the business toward fewer, higher-value jobs.

conversion 2×+handles a large share of jobsavg order value up materially
pricing engineNext.jsStripeAPI integration
/04

Automation & integration backbone

Led architecture; built hands-on

An idempotent, rate-limit-aware integration layer connecting CRM, field-service, finance and e-sign. Replaced a fragmented, multi-system manual process with automated gate-based tracking and a single live control panel — removing manual data entry and month-end reconciliation.

≈ half of activity automated~200 staff-hours/month saveddispatch 12 → 3 min
n8nwebhooksCloudflarePostgresworkflow automation
/05

Web platform & growth engineering

Led & delivered

Migrated multiple brand sites to Astro + Cloudflare with zero loss of search rankings, consent-gated analytics and CRM lead capture; ran technical SEO and Core Web Vitals programmes that grew organic demand while cutting acquisition cost.

no ranking loss on migrationorganic demand up double-digitsCAC down
AstroCloudflaretechnical SEOperformance
02

How I build

The discipline behind shipping AI that a business can actually rely on.

// 01

Evals, not vibes

LLM features ship against golden-set evals, regression checks on prompt/model changes, and citation-grounding gates — so reliability is measured, not assumed.

// 02

Tied to the P&L

I start from the business outcome and work back to the system. Every build has a number it's meant to move — and I own whether it moved.

// 03

Systems over heroics

I take businesses from owner-dependent to system-led: documented processes, automation and dashboards so things run without depending on any one person.

03

Stack

Hands-on across the whole path — model to production.

AI / LLM
RAGLLM agentsPrompt engineeringEmbeddingsEvals & groundingClaude / OpenAI / Gemini APIs
Languages
TypeScriptPythonJavaScriptSQL
Build
NodeNext.js / ReactAstroTailwindPostgres / pgvector
Cloud / Infra
Cloudflare (Workers · Pages · KV · R2)VercelCI/CDmonitoring
Automation
n8nREST / webhooksCRM & finance integrationPower BI
04

The short version

A deliberate climb from creative production to building the systems a business runs on.

2017 — 2021
Broadcast camera & production
Camera work on BBC & ITV productions. Where I learned to ship to a deadline, to a standard, with a team.
2021 — 2023
In-house video & content
Built a services group's content capability — then started automating my own job, and didn't stop.
2023 — 2025
Technical & systems lead
Self-taught into software, AI and systems architecture. Unified the company's platforms, led a small dev squad, shipped the booking and automation stack.
2025 — now
Running the business · building its AI
Operate the company day-to-day while personally building the LLM platform, decision tools and integration backbone it depends on.