Matthew has worked with British Airways during his time at AND Digital to help maintain and deliver new features to their app. Particularly focusing on functionality to improve customer experience during COVID times.

A sophisticated cryptocurrency tracking and portfolio management app built with Kotlin Multiplatform, delivering a native experience on both iOS and Android from a single codebase. Features real-time market data via WebSocket connections, professional-grade interactive charts with technical indicators (SMA, EMA, RSI, MACD), multi-portfolio management with P&L calculations, and smart price alerts with push notifications.

Built using Kotlin Multiplatform (70%+ code sharing), Jetpack Compose & SwiftUI for native UI, SQLDelight for persistence, Ktor for networking, and Clean Architecture with MVVM. Includes custom high-performance chart rendering engine, WebSocket connection pooling, and background service implementation for both platforms while maintaining 60fps performance.

The XLN sales app is an android app developed for sales agents at XLN. The app shows sales agents prospective clients in their area, allows an agent to assess which prospective clients are most valuable and/or most likely to have a lead. It also allows agents to edit and upload contracts on the go. Built using android and firebase.

Volcano Escape is a web-based game built with three.js. The aim of the game is to escape from the volcano before the ever-rising lava swallows you up.

The "Mehmet" app is a meetings app. Its primary use case is to manage when/if participants are available for a meeting and then resolve conflicts. Built using Android and Firebase.

Paint is a drawing application written from scratch in C++.

An application written in Clojure which features a maze generator and a solver. It works on a client server based model. The client can request mazes of different sizes and request solved versions. This maze/s are presented to the user using an eye pleasing GUI.

Plain english a an esoteric programming language which attempts to be similar to natural language. It is written using ANTLR and compiles to JavaScript.

The lyrics generator works by taking input on the first verse or verses of a song and it's genre. It then tries to generate a whole song based on this input. The lyrics were scraped from azlyrics.com. We used multiple different models, BERT and GPT-2(fine-tuned and vanilla). We also implemented rhyming. We then took these different implementations and compared and evaluated them.

The goal of this project was to create a system that could profitably trade different securities(stocks, commodities, forex, cryptocurrencies etc). I looked at four different implementations, three of which had already been proposed in academic papers and one which was my own implementation using machine learning. The project was built using Python and Tensorflow.

This project includes a full suite of tools needed to run and maintain this crypto currency trading bot. The goal of the project is to implement a crypto currency trading strategy that I created and to have it run 24/7. It features a special interface to monitor market conditions on different exchanges. It has a module to run the strategy, which continuously monitors the market for trading opportunities and places orders depending on a set of conditions. And finally is a integration with Google Stackdriver and also status updates forwarded to a Telegram channel.

The restaurant management system project is a project I originally made for a friend. He owned a restaurant but had a problem, working out how to set up the tables and where to place customers in a way that was optimal was very time consuming, therefore I created this application for him. The application works across desktop and mobile, built on nodejs, vue and dynamodb. I'm currently working on scaling the application to multiple clients.