Selected work

Three systems, built end to end

Each one went from an empty repository to something that runs. Drag the cards sideways for the short version, or read the detail underneath.

Projects 3 Domains fintech, real time, commerce

Abstract illustration of a microservices financial application

Microservices architecture

Enterprise Financial Application

A financial mobile application split into services that deploy on their own. Spring Boot backend, Docker images, and a Jenkins and GitLab pipeline that built and tested every merge before anyone saw it.

Independent services, one pipeline

Real-Time Chat & Trivia

Two mobile applications on one real-time spine: messages and game state stay in sync across every player without a refresh. The API suite was automated in Postman.

30% less manual verification

GoStore

A storefront plus the part customers never see: C# WinForms dashboards where staff manage catalogue, stock and orders, on a SQL data layer modeled for real reporting.

Storefront and back office, one system

Everything else is on GitHub

Repositories, commit history, and the work that did not make this shortlist.

Open GitHub

Drag, or use the arrows

In detail

Role, stack, outcome

Microservices

Enterprise Financial Application

Backend & delivery

A financial mobile application built as independent services rather than one block, so a change to one part could be released without redeploying the rest. I wrote the Spring Boot backend and packaged each service as a Docker image.

Jenkins and GitLab ran the pipeline: every merge built, tested and produced an artifact, which meant a broken change was caught by the machine rather than by a person on release day. Screens were designed in Figma and agreed before implementation.

My part

Backend services, containers, CI/CD

Stack

Java, Spring Boot, Docker, Jenkins, GitLab, Figma

Outcome

Services release independently on a repeatable pipeline

Real time

Real-Time Chat & Trivia

Full stack & test automation

Two React Native applications sharing one problem: many people, one state, no refresh button. Messages arrive as they are sent and a trivia round stays in step for every player, which puts the difficulty in the connection layer rather than the screens.

I automated the API suite in Postman so the checks that ran before every release stopped being typed by hand. That removed about 30% of the manual verification round and moved attention to what had actually changed.

My part

Client, server, automated API suite

Stack

React Native, Expo, Node.js, Postman

Outcome

30% less manual verification per release

Commerce

GoStore

Full stack

An e-commerce platform with both halves built: the storefront customers use, and the C# WinForms dashboards staff use to manage catalogue, stock and orders.

The data layer was modeled in SQL around the questions the business kept asking — what sold, what is running out, which order is stuck — so answering them was a query rather than a spreadsheet export.

My part

Storefront, dashboards, data model

Stack

C#, WinForms, SQL Server

Outcome

Staff manage the shop without touching the database