AI & Full-Stack Engineer · Devon, UK · Available for freelance

I build AI products
end-to-end — and ship them live.

Generative-AI pipelines, real-time web apps, payment backends, and self-hosted infrastructure. Designed, deployed, and secured solo — from a prototype to a working, paying product.

Live product: ai3d.pro Stack: Python · FastAPI · WebGL · Docker · Linux Specialism: Generative AI & ML
What I do

Services

I cover the whole stack, so you don't have to stitch five contractors together. My focus is generative AI — but I build, deploy, and secure everything around it too.

🧠

Generative AI & ML Specialism

Stable Diffusion & SD Forge deployment (including the tricky AMD ROCm path), AI image-upscaling pipelines, 3D reconstruction & Gaussian Splatting, and clean Gradio web apps that put your model in front of non-technical users.

Stable DiffusionROCm Gaussian SplattingGradio PyTorch
🌐

Full-Stack Web & Backend

FastAPI / Python backends, Stripe & PayPal payments with server-enforced sessions, REST APIs and auth, plus real-time WebGL / Three.js front-ends and audio-reactive visuals.

FastAPIStripe PayPalThree.js REST / JWT
🖥️

DevOps & Self-Hosting

Docker & Compose, nginx, Linux/VPS administration, Nextcloud · Mailcow · Pi-hole · Kavita, TLS & DNS automation — and proper security hardening (ufw, fail2ban, audits). This very site runs on infrastructure I manage.

Dockernginx LinuxLet's Encrypt Hardening
🔌

Embedded & IoT

ESP32 firmware, RF / SDR projects, and Home Assistant integrations — from streaming a doorbell feed to a custom display to decoding a thermostat's 433 MHz signalling.

ESP32Arduino RTL-SDRHome Assistant
Proof, not promises

Selected work

A sample of real, shipped projects. Every one was built, deployed and secured hands-on.

⭐ Flagship · Live product

ai3d.pro — a paid AI music visualizer, built solo

A real-time WebGL / Three.js engine driven by live audio (FFT + beat & BPM detection), backed by a FastAPI payment service with Stripe & PayPal, nginx access-gating, self-hosted GeoIP analytics with a live 3D globe, full SEO, and a security-hardened VPS.

Idea → live, HTTPS, payment-enforced product. End-to-end, solo.   Visit ai3d.pro →
Generative AI

SD & Gaussian Splatting on AMD/ROCm

Stable Diffusion Forge + an OpenSplat 3D-reconstruction pipeline running on a consumer AMD RX 7800 XT — the non-NVIDIA path most guides skip.

ML on cost-effective hardware.
Infrastructure

A self-hosted homelab

Nextcloud, Mailcow email, Pi-hole, Kavita and nested KVM VMs across Raspberry Pis and a VPS — Docker, nginx, TLS and hardening throughout.

Private SaaS replaced, fully self-run.
Embedded / IoT

Custom hardware integrations

ESP32 firmware streaming a doorbell's live feed to a display, and a 433 MHz RF sniffer reverse-engineering a thermostat — bespoke work you can't buy off the shelf.

Firmware → working device.
Systems software

C engine work & native apps

Real changes to the EDuke32 C game engine (GTK loader, custom UI, a Wayland crash fix) plus PyQt5 + OpenGL desktop apps.

Low-level to GUI, comfortably.
How I work

Clear, tested, handed over

01

Scope clearly

We agree exactly what success looks like before any code is written.

02

Check in early

I confirm before making big assumptions — no nasty surprises at delivery.

03

Verify by running it

I don't hand over code I haven't watched work. Everything is tested live.

04

Secure & document

You get a hardened, documented, maintainable result you can run yourself.

Let's build something

Have a project in mind?

Whether it's a generative-AI build, a web app with payments, or self-hosted infrastructure — tell me what you're making and I'll tell you honestly how I'd ship it.

Email me →