Data has always felt like a puzzle to me, and AI is the most powerful tool we have to solve it. I'm currently pursuing my MSc in Artificial Intelligence, where I've been deepening my expertise in machine learning, data engineering, and predictive analytics. I love the moment when raw numbers transform into something meaningful — a model that predicts, a dashboard that guides, a decision that sticks.
I'm building toward a career where I can apply cutting-edge AI to real-world data challenges.
When I'm not wrangling data, you'll find me exploring new analytical techniques, contributing to open-source tools.
Breaking down complex problems into clear, structured questions
Building and applying intelligent models that learn from data
Building reliable pipelines that scale with business needs
Crafting dashboards and reports that speak to any audience
Translating technical findings into business recommendations
Hands-on work that showcases end-to-end data skills. Explore my work!
An early diabetes prediction system developed using Python, Scikit-learn, and Streamlit to explore how machine learning can support healthcare decision-making.
Built as part of my AI learning journey, the project combines exploratory data analysis, predictive modelling, and a Streamlit-based web application to demonstrate how AI can be applied to real-world healthcare challenges.
An end-to-end healthcare analytics pipeline developed using SQL and MATLAB to predict worsening asthma symptoms using synthetic NHS-style patient data.
This project applies the CRISP-DM methodology to perform data engineering, feature engineering, predictive modelling, and deployment architecture design for large-scale healthcare analytics.
The project explores how Transformer-based models can capture sentiment and context within large-scale user-generated text data.
The project covers data analysis, Transformer tokenisation, model fine-tuning, and performance evaluation while demonstrating how modern AI techniques can transform raw player feedback into meaningful insights.
A real-time facial recognition system built with Python, OpenCV, and Streamlit to explore how dual-model verification improves reliability under live, unpredictable conditions.
This project combines RetinaFace detection, parallel embedding extraction with FaceNet and ArcFace, cosine-distance identity matching, and a live evaluation module for per-model accuracy.
The foundations behind the work.
Whether you have a project in mind, a role to discuss, or just want to chat about data — my inbox is always open.