Hi there, I'm Muhammed :)

Spam/Ham Detector

A machine learning web app that detects whether a text message is spam or legitimate (ham). Built with a Support Vector Machine trained on real SMS data, using TF-IDF vectorisation to understand language patterns.

AI-Driven Defect Detection in Composite Materials

MSc Dissertation, Northumbria University 2025

Designed a semi-supervised pipeline combining K-Means clustering and YOLOv8 to reduce manual labelling in industrial defect inspection. Validated pseudo-labelling quality using Silhouette and Davies-Bouldin scores, built a YOLOv8 detection baseline, and documented trade-offs between annotation cost, accuracy, and deployment feasibility.

Student Placement Platform

MSc Group Project, Northumbria University 2025

Delivered a full system specification for a student placement platform within a five-member team. Translated stakeholder needs into functional and non-functional requirements, produced UML models (use case, activity, class diagrams) ready for developer handover, and conducted a critical reflection covering technical constraints and ethical considerations.

Nim Game

Northumbria University 2025

A Java implementation of the Nim strategy game with both a Swing GUI and text-based interface. Used the Observer pattern to decouple game logic from the UI, and extended gameplay with AI opponents, undo, and save/load features. Validated with a full JUnit test suite covering black-box, white-box, and boundary cases.

FPL Recommender System

Northumbria University 2025

Designed an intelligent UI for a Fantasy Premier League recommender system, grounded in user research. Prototyped data-driven screens surfacing ML recommendations with visual explanations, and evaluated fairness, transparency, and trust in line with responsible AI principles.