Full-stackFull-stack

engineer.engineer.

Problem

A lot of people in Nigeria still use feature phones, or have a smartphone but no data. They can't use a normal crypto wallet app. I wanted to see if you could put one behind a USSD code instead — the same way mobile money already works.

Approach

A Go backend that talks to telco USSD gateways and translates *123#-style menus into Stellar transactions. The phone is dumb; session state, keys, and rate-limiting all live on the server.

Outcome

An early prototype that proves the idea works end-to-end. A lot more to harden before it's production-ready, but the core flow is there.

Problem

Hospital ↔ HMO settlement in Nigeria is slow and full of disputes. I was curious whether the Stellar transaction model — cheap, fast, deterministic — could be a fit.

Approach

A minimal Rust service that orchestrates provider-payer reconciliation and tokenizes claim instruments on Stellar. Still a sketch, not a product.

Outcome

Useful as a learning exercise in domain modelling and tokenized state. Helped me get comfortable with Stellar's account model.

Problem

I wanted a project that touched a bunch of things at once — auth, persistence, transactions, money math, validation. A wallet was the perfect excuse.

Approach

Backend in Python with a clean separation of accounts, transactions, and budgets. Added simple budgeting tools, transaction history, and rewards tracking on top.

Outcome

Solidified the basics of API design, money handling (never floats), and writing tests that actually catch off-by-one bugs.

Problem

I wanted a project that actually called external APIs and handled the messy real-world data that comes back.

Approach

React frontend that talks to The Movie Database (TMDb) API. Search, filtering, detail pages, the basics.

Outcome

Got more comfortable with React state, debouncing, pagination, and dealing with APIs that don't always return what you expected.