I build and deliver end-to-end analytical projects that solve real business problems — combining machine learning, deep learning and LLMs with clear business storytelling. Currently open to Data Analyst, Data Scientist and Business Intelligence roles.

• 🎓 Master's in Business Analytics from the University of Leeds
• 💼 Background in Data Analytics + Business Development — I speak both technical and commercial
• 🤖 Currently building: Real-time Fraud Detection with XGBoost, LSTM and GPT-4o-mini
• 🌱 Learning: LLM integration, MLOps, cloud deployment
• 📍 Based in the UK | Open to remote and hybrid roles

Languages

Machine Learning & AI

Data & Visualisation

Tools & Platforms

Production-grade fraud detection pipeline on 590K real transactions combining three complementary ML approaches with LLM-powered explanations.

• Built ensemble pipeline: Isolation Forest (unsupervised) + XGBoost (AUC 0.924) + LSTM sequence model (AUC 0.891) to catch fraud patterns no single model detects alone
• Engineered 12 domain-specific fraud features including transaction velocity, card spend deviation and time-based anomaly flags
• Integrated SHAP + GPT-4o-mini to auto-generate plain English fraud analyst reports for every flagged transaction
• Deployed as a live Streamlit dashboard with real-time transaction simulation, fraud alerts and model comparison visualisations

Python XGBoost TensorFlow LSTM SHAP OpenAI Streamlit Plotly

End-to-end churn prediction pipeline for a telecom company — from raw data to deployed ML model to Power BI executive dashboard.

• Cleaned and analysed 7,043 customer records across 21 features, identifying key churn drivers including contract type, tenure and monthly charges
• Built and compared three models — Logistic Regression, Random Forest and XGBoost (AUC 0.91) — with full SHAP feature importance analysis
• Wrote advanced SQL queries (window functions, CTEs, CASE WHEN aggregations) to segment customers by risk profile
• Delivered a Power BI dashboard with slicers by contract type, risk segment and internet service for business stakeholder reporting

Python pandas XGBoost SQLite SQL Power BI scikit-learn seaborn

Quantitative investment model using time series analysis and ML to forecast FTSE 100 stock returns.

• Analysed 500K+ data points using ARIMA, GARCH and ensemble ML models
• Built in R with full statistical validation and backtesting framework
• Delivered actionable investment signals with confidence intervals

R ARIMA GARCH Time Series Python scikit-learn

ML platform for cryptocurrency ICO investment analysis using NLP and sentiment analysis.

• Built predictive models using Naive Bayes classification on ICO whitepaper text data
• Implemented end-to-end sentiment analysis pipeline on social and news data
• Combined NLP signals with financial features for investment scoring

Python NLP Naive Bayes Sentiment Analysis Machine Learning

Interactive executive dashboard analysing £963M revenue across 22 countries for strategic decision making.

• Identified market expansion opportunities and underperforming regions through geographic revenue analysis
• Built drill-through reports and dynamic slicers for C-suite stakeholder consumption
• Connected Python data pipeline to Power BI for automated refresh

Power BI Python SQL DAX Data Modelling

💼 Open to: Data Analyst · Data Scientist · Business Intelligence · ML Engineer roles

🤝 Available for: Full-time · Contract · Remote · Hybrid