Software that doesn’t fall over.
I build the backends, payments and permissions products run on, and the AI automations that take repetitive work off a team. Every system ships with its own paper trail: auth model, access rules and data shape, documented and testable.
My preference is boring in the useful sense: settle who can access what, how data behaves under pressure, and what happens when something fails, before adding more surface area.
hi, I'm Henry
I’m a full-stack engineer in Lagos. I build the parts of a product people rely on every day (sign-in, permissions, payments, inventory) and AI automations that take repetitive work off a team: finding and qualifying leads, turning an idea into a researched article, pulling data out of invoices, writing the weekly report. Right now that’s a poultry platform used by 30+ farms, a study app students open daily, and a bill-splitting app on both app stores.
How I work
Evidence before atmosphere
Screenshots, numbers and logs before adjectives.
Tools enforce, prompts suggest
If an agent must never send an email, it doesn’t get a send tool.
Done means the real UI drove the real backend
Green tests aren’t done. Clicking through the real app once is.
Every failure names who fixes it
An error that says which setting to change beats a silent retry.
Software engineering
Backend-leaning full-stack. The consequential parts, built to survive production.
- Backend systems & APIs01
- Auth, RBAC & permissions02
- Billing, ledgers & audit trails03
- Production reliability & CI/CD04
- Mobile (Expo) & web (Next.js)05
AI automation
Agents and workflows that do real work and still keep a person in charge.
- Claude Agent SDK agents01
- n8n workflows02
- Document extraction & OCR03
- Evaluation, guardrails & human review04
- Cost caps & run logging05
Featured work
Customer Support Agent
A voice and chat support agent for a payments company. It answers from the knowledge base, checks a payout only for its owner, and hands real cases to a person. It never moves money.
Lead Research Agent
Turns “find 10 B2B SaaS companies that need automation” into qualified leads with evidence, and review-ready outreach it can never send itself.
Aviarai
A poultry farm’s whole day (eggs, feed, dispatch, billing) running through one system used by 30+ farms and 70+ people.
Divyd
From “who owes who” to a group that splits bills and actually talks, live on iOS and Android.
Every automation makes the same five stops.
Trigger
what starts it
Gather
where the facts come from
Think
what reads and writes
Check
what stops it being wrong
the stop most people skipDeliver
where results go
Toolbox
The ones with a dot shipped in a project above. Point at one to see which.
Commits
One square per day for the last year. Most of it lives in private client repos.