An AI learning platform that turns any topic into a structured course.
Curates real videos, docs and Q&A instead of making lessons up. Multi-stage LLM pipeline with reranking. Built solo.
Came for the Master's. Stayed for the Spoons.
Software Engineer · AI/ML · LLM systems
I was 16. School computer lab. C++.
First time something did exactly what I told it to.
For the first time, it was my world, and I was in control.
From the Blue Line to the Yellow Line. Keep scrolling, we're moving →
First C++ program. Fell for it immediately.
MERN, Next.js, IoT, ML. Google Developers Club. Showed an IoT agri-tech project at an IEEE exhibition.
Python scraping pipelines across 5 platforms. 10,000+ profiles collected and queried.
15+ e-commerce sites on WordPress & Shopify. Kept 20+ live sites alive.
12+ full-stack apps for 8+ clients. Multi-vendor stores with 5,000+ listings. ERPs that cut data entry 60%.
Started 5 or 6 months before my Master's even began. Hit the ground running.
Did it while working. Deep Learning, ML, Secure Software. Accenture hackathon: commended for architecture.
LLMs, RAG, MCP servers, Kubernetes, 55 VMs and the on-call rotation.
Dropped my bags at student accommodation and walked into the first place I saw.

This service calls at Haymarket, Monument, Central, Gateshead and Gateshead Stadium.
Please mind the gap between the project and the README.
Where my MSc lived. Also where most of my sleep went.
Where my MSc lived. Also where most of my sleep went.
An intrusion detection & prevention system that stops attacks at the edge and thinks about them in the middle.
Side projects, posted where every good idea at uni ends up: the Union noticeboard. Tear off a tab to see the code.
An AI learning platform that turns any topic into a structured course.
Curates real videos, docs and Q&A instead of making lessons up. Multi-stage LLM pipeline with reranking. Built solo.
Intrusion detection & prevention that blocks at the edge and thinks in the middle.
eBPF/XDP agents on edge devices, a central LLM reasoning engine in Go, Kafka telemetry. My MSc dissertation (Distinction).
A job-application tracker that gets progressively crueller until I apply for something.
Never typed into: it reads my inbox, parses 93% of emails with plain code and hands the weird 7% to an LLM. No nagging, just arithmetic.
Turns my bank account into one number: what I can spend each day until payday.
Reads Monzo (read-only, structurally), finds payday on its own, and lets an overspent budget borrow, visibly.
My server's file store, with a map of every byte and what's gone cold.
A treemap where area really is bytes, a COLD mode for dead weight, and an inbox where an LLM suggests filenames (it only ever sees the name).
Hiring, building something, or need a Spoons recommendation? My inbox is open, usually after midnight.