From interfaces to integrations to intelligent systems.
My career has taken me from building the software people interact with, to connecting systems behind the scenes, and now to building AI that can reason across them. Today, I'm focused on agentic systems that turn intelligence into action.
Frontend → Full-stack → Integration → Agentic AI
I build systems that connect language models to software, data, and real-world workflows. Frontend, full-stack, and integration work shaped how I think about intelligent systems: not as isolated chat interfaces, but as software that has to use tools, hold context, and operate reliably.
commit c481d0f (HEAD)
2026 → present
ScottishPower
Energy
Agentic AI Developer
Now building agents that reason over data and coordinate action across systems.
Stack: LangGraph, RAG, Multi-agent
commit a19f42e
2025 — 2026
Social Security Scotland
Government
Integration Developer
Shifted from building interfaces to connecting systems: event-driven integrations across cloud services.
Stack: AWS Lambda, SQS, API Gateway
commit 7d3e8b1
2024 — 2025
Big Bite
Digital Agency
Software Engineer
Moved deeper into the stack, shipping full-stack product work end to end.
Stack: React, Node.js, Full-stack
commit 2f6a19c
2020 — 2023
OPay / Redbracket
Fintech
Frontend Engineer
Built the interfaces people touch: React apps wired to REST APIs, shipped across markets and languages.
Stack: React, i18n, REST APIs
The common thread: building software that connects things together.
How I think about the work
AI that can actually do something
The interesting part of an LLM isn't what it can say. It's what happens when it can use tools, hold context, and act on real systems.
Systems, not demos
A prototype can produce an impressive response. A production system has to handle failure, ambiguity, state, and permissions.
Connecting the pieces
My work has always been about connecting things: interfaces to APIs, services to services, and now language models to the software around them.
Engineering over hype
I care about where AI actually removes friction, not about chasing whatever's newest.
Things I've built outside the day job
EdusythPro
Synthesizes educational content into engaging, efficient material for students and educators.
Explored: LLM integration, prompt design, content transformation
Shipaid Claims Portal
LiveA claims portal that lets users file claims for lost or damaged packages.
Explored: Remix data loaders, GraphQL APIs, production form flows
Speeco
LiveA speech-to-text notes app, originally built for my undergraduate thesis, since rebuilt on Vue 3 with an on-device Whisper fallback for browsers without native speech recognition.
Explored: Browser speech APIs, on-device ML fallback, state management
Crecon
A testing playground built during a dev training session: Cypress and Jest against a Vue.js app.
Explored: End-to-end and unit testing patterns, test-driven workflows
Would You Rather
A full React app with Redux, letting users play the classic "Would You Rather" game.
Explored: Redux state architecture, component styling patterns
What I reach for
agentic-ai: Building systems around LLMs: retrieval, orchestration, multi-agent workflows, and tool use. Python, LangChain, LangGraph, AutoGen, Semantic-Kernel, RAG, Vector-DBs
cloud-integration: Connecting software across APIs, messaging, and enterprise platforms. AWS-Lambda, Amazon-SQS, API-Gateway, Boomi, Kibana
full-stack: The foundation everything else runs on top of. TypeScript, React, Next.js, Node.js, Java, SQL