Full-stack engineer working across backend systems, APIs, and AI-integrated products. Co-founder of WEBROS. I've shipped six live products — every one of them written on an Android phone.
Core technologies
From the database to the pixel — I take products from idea to launch, with engineering rigour and an eye for design.
End-to-end builds: architecture, backend, frontend, and deployment — owned from first commit to launch.
Robust server-side systems built for real traffic — from API design to authentication, rate limiting, and deployment.
Bringing LLMs and AI assistants into real products — from prompt engineering to production features users rely on.
Responsive, accessible interfaces in React and modern JS — fast-loading and faithful to the design.
Data architecture that scales — relational and NoSQL, designed for query performance, consistency, and growth.
Sites that load instantly and get found — semantic markup, structured data, and Lighthouse-tuned builds.
A selection of platforms I've designed, built, and launched — most are live right now.
An AI-powered study companion for Nigerian students preparing for WAEC, JAMB, NECO and Post-UTME. Summarises notes, generates quizzes, builds revision plans, analyses YouTube lectures, runs a community past questions engine, and works offline. Powered by Vexio — a built-in AI tutor that explains any topic or exam answer in full detail.
Visit live siteA GitHub profile analyzer — drop in any username and get their languages, repos, activity, and a developer score in one clean view. Built with no backend, just vanilla JS and the GitHub API.
Visit live siteAn AI prompt engineering tool that turns any idea into five levels of expert prompts — from quick starter to enterprise-grade master — across categories like development, business strategy, content creation, AI art, and local Nigerian business. Supports multiple tones and an enhanced mode for deeper outputs.
Visit live siteA campus platform connecting students to their academic world — architected and built from the ground up. Full case study and live link coming once it ships.
Case study soonA browser-based tool that reads any PDF or pasted text aloud, highlighting each word as it's spoken. Supports 15+ languages with automatic language detection, adjustable speed, pitch and volume controls, and audio export as a .webm file. All processing happens on-device — nothing is uploaded.
Visit live siteAn AI-powered personal workspace — chat with an AI assistant, manage tasks, capture notes, track habits, and organise your work in one place. Built as a fully installable Progressive Web App with persistent local storage. The same Vexio AI that powers Aveliq, now as a standalone productivity tool.
Visit live siteAlongside running my own builds, I co-founded WEBROS — a lean web development studio shipping modern, AI-powered products for founders and businesses across Africa and beyond.
I build for outcomes, not just features — every decision ties back to the user and the goal.
Lean iterations and clean code — real products in users' hands quickly, without cutting corners.
Async-friendly, reliable, and easy to work with — on time and on brief, every time.
I'm George Erubami, a software engineer and co-founder of WEBROS, based in Ibadan, Nigeria. Here's the part most people don't expect: I don't own a laptop. Everything I build — backend services, REST APIs, React frontends, LLM integrations, the whole stack — I write on an Android phone, in Acode, deployed through Termux. It's slower than a real dev setup in some ways and it's forced me to think harder about architecture before I touch a single line of code, because there's no room for sloppy iteration when every edit costs you something.
I taught myself by shipping. Six live products so far, real people using them, real bugs I've had to fix in production. That's still how I learn fastest.
"I didn't wait for someone to teach me. I built my way to competence — one shipped product at a time."
Short notes on engineering, products, and lessons from shipping real things.
The real engineering decisions behind Aveliq — from prompt design to Firebase architecture and why I chose Node.js over serverless.
How I built a clean GitHub profile analyzer using pure vanilla JS — no framework, no backend, just the GitHub API and a focused UI.
Two years of building in public, shipping products, and making mistakes — the honest version of what it actually takes to go from zero to full-stack.
Hiring, collaborating, or starting a project? Send a note — I reply within a day.