Projects

• Software Development

  Dynamic eLearning Program Powered by AI
  

  

  Description

  For my major university project, I developed an innovative, scalable eLearning platform designed to integrate seamlessly with existing business infrastructures. Built using the .NET framework with C# for backend development and SQL for retrieving, processing, and storing data into an existing database, the system was created to ease the transition for businesses already using eLearning by allowing full integration of their current employee databases. The platform leverages the LLaMA 3.1 8B Large Language Model, applying parameters such as an employee’s role, industry, and experience level to dynamically generate real-world scenarios and adaptive, AI-driven assessments tailored to specific job functions. This ensures highly relevant, engaging learning experiences that improve both knowledge retention and practical application. Designed with enterprise flexibility and scalability in mind, the platform supports diverse organisational structures and training workflows, making it suitable for both independent businesses and large enterprises seeking a more intelligent and personalised approach to workforce development.

  Technologies

  

  Shift and Rota Management Program
  

  

  Description

  A management tool was created for a local veterinary practice, identifying the key requirements and functionalities needed for the program. This included tracking employee schedules, managing shift changes, and assigning tasks. I used C# in the .NET framework for the front-end development, providing a user-friendly interface and easy navigation for employees and managers. For the backend, I used SQL to manage the database and retrieve specific scheduling information, designing a schema to include tables for employees and shifts. The program connected to a backend relational database to store and manage all necessary information, including employee details, shift schedules, task assignments, and rota information. Overall, the use of technologies such as C# and SQL in the .NET framework enabled the creation of a comprehensive management tool for employee shifts and rota information, streamlining scheduling processes and improving overall efficiency.

  Technologies

  
• Web Development

  Clearwater Accountancy Website
  

  

  Description

  For my first professional project, I independently developed a website for an accountancy firm, leveraging my skills in HTML5, CSS3, and JavaScript. I incorporated external APIs to enhance specific processes and tailor the website to the project's scale. To ensure data security and encryption, I implemented SSL certificates. Through thorough research, I carefully selected and used only necessary resources to optimise the website's performance and efficiency. This project not only strengthened my existing skill set but also allowed me to effectively deliver a functional and secure website for the accountancy firm in an area where data security is most valuable.

  Technologies

  

  Gaming News Website
  

  

  Description

  I have embarked on an ongoing journey to create my own news article website focused on gaming, aimed at enhancing my web development skill set. This project has been instrumental in deepening my understanding of backend web development, particularly with HTTP protocols and PHP. Additionally, it has allowed me to develop advanced scripting capabilities, significantly broadening my technical expertise. Through this endeavor, I am not only building a platform to share gaming news but also honing my abilities in creating robust, dynamic web applications. This project has become the foundation of my advanced web development skills, while I have actively attempted to create each component to my full potential with optimal web development practices, such as responsive design that does not require breakpoints/media queries and functions flawlessly.

  Technologies

  
• Scripting

  Digit Recognition Program
  

  

  Description

  I developed an AI-powered script using Python, utilising a large dataset and handwriting characteristics to detect handwritten numbers with high accuracy. The script employed machine learning algorithms, specifically convolutional neural networks (CNNs), to analyse the digitised images and identify the corresponding numbers. The algorithm was trained on a vast dataset of handwritten numbers, enabling it to recognise and learn the distinct characteristics of each digit, such as shape, size, and stroke patterns. Additionally, the script calculated the probability and confidence of each detection, providing a comprehensive output. This allowed for accurate analysis and verification of the results, as well as identification of potential errors or ambiguities. Once the script finished processing, it displayed the detected numbers, along with their corresponding probabilities and confidence levels, enabling users to make informed decisions.

  Technologies

  

  Predicting House Sale Prices
  

  

  Description

  In this project, I developed a machine learning model to predict house sale prices based on various property features and location data. I began by thoroughly evaluating the dataset and the nature of the prediction task, ultimately identifying it as a regression problem—one that requires a supervised learning approach to estimate continuous numerical values. After preprocessing the data, handling missing values, and performing feature engineering to improve model performance, I implemented and tested three regression algorithms: Linear Regression, Random Forest, and XGBoost. Each model was evaluated using standard regression metrics such as R² score, Mean Absolute Error (MAE), and Root Mean Squared Error (RMSE) to assess prediction accuracy and generalisation. The goal was to identify which model delivered the highest predictive capability based on the dataset's characteristics. Ultimately, the results demonstrated that XGBoost provided the best overall performance, outperforming both Random Forest and Linear Regression in terms of accuracy and robustness. This project showcases my ability to select appropriate machine learning algorithms, perform model evaluation and comparison, and build effective solutions for real-world regression problems using supervised learning techniques.

  Technologies

  
• Data Warehousing

  Bookstore Chain Data Warehouse
  

  

  Description

  In this project, I designed and implemented a fully functional data warehousing system for a proposed multi-location bookstore chain using PostgreSQL. The goal was to provide a scalable infrastructure capable of integrating, organising, and analysing data from various branches to support business intelligence and strategic decision-making. I began by creating a robust ETL (Extract, Transform, Load) pipeline that streamlined the ingestion of sales, inventory, and customer data from multiple sources into a central data warehouse. To ensure high performance and scalability, I utilized advanced PostgreSQL features such as indexing for faster query execution and TimescaleDB partitioning to efficiently manage and query time-series data across different store locations. Once the data was structured and optimized, I developed a suite of advanced SQL queries to extract meaningful insights into key business metrics. These queries allowed stakeholders to evaluate store performance by location, analyse sales trends over time, and identify top-selling products and categories. Additionally, customer data was leveraged to enable targeted promotional campaigns, by identifying individuals most likely to benefit from specific offers based on their purchasing behavior. This project demonstrates my ability to architect a complete data warehousing solution, optimise data storage and retrieval, and deliver actionable insights through intelligent querying and analytics. It reflects my proficiency in database design, performance tuning, and the practical application of data engineering techniques to support data-driven decision-making in a commercial context.

  Technologies

  

What's next?