BSc Data Science student at KCA University | Expected graduation: September 2027
Focused on Data Science, Machine Learning, Business Intelligence, Applied AI, and Generative AI. I develop practical, data-driven applications that transform raw data into insights, predictions, and intelligent solutions.
- Programming and data: Python, SQL, SQLite, pandas, NumPy, Excel
- Machine learning and explainability: scikit-learn, Logistic Regression, Random Forest, GridSearchCV, SHAP, feature engineering, cross-validation, model evaluation
- Data analytics and visualization: Exploratory Data Analysis (EDA), Business Intelligence, Matplotlib, Seaborn, Plotly, Recharts
- Applied AI: Rule-based NLP, Retrieval-Augmented Generation (RAG), semantic search, vector databases, Generative AI
- Web development: React, TypeScript, JavaScript, FastAPI, HTML, CSS, Tailwind CSS, Vite
- Development and deployment: Git, GitHub, VS Code, Jupyter Notebook, Streamlit, pytest, Cloudflare Pages, Render
A deployed full-stack business intelligence platform designed to analyze retail and e-commerce performance. Processes approximately 1.03 million retail transactions and presents interactive analytics across six dashboards: Overview, Sales, Products, Customers, Geography, and Returns.
Features KPI monitoring, interactive charts, data filtering, and REST API-powered analytics.
Tech: React, TypeScript, Tailwind CSS, Recharts, Python, FastAPI, SQLite, Cloudflare Pages, Render
Explainable financial transaction fraud detection using Random Forest and SHAP, built with Python and Streamlit. Supports single-transaction and batch predictions, with 47 passing automated tests.
Tech: Python, pandas, NumPy, scikit-learn, Random Forest, SHAP, Streamlit, pytest
Final held-out evaluation:
| Metric | Result |
|---|---|
| Precision | 0.945946 |
| Recall | 0.736842 |
| F1 | 0.828402 |
| PR-AUC | 0.809520 |
These metrics come from the final historical holdout evaluation. Synthetic demonstration transactions are used only to demonstrate the interface and are excluded from these evaluation results.
Customer churn classification developed during the Lloyds Banking Group Data Science Job Simulation. Covers data preprocessing, exploratory analysis, feature engineering, Logistic Regression, hyperparameter tuning with GridSearchCV, cross-validation, and model evaluation.
Tech: Python, pandas, scikit-learn, Matplotlib, Seaborn
Rule-based Python chatbot developed during the BCG GenAI Job Simulation. Answers predefined financial questions using structured financial data and pandas analysis.
Tech: Python, pandas, rule-based NLP, financial analysis
Python and SQLite inventory management system supporting authentication, product management, stock tracking, sales recording, reporting, and persistent database storage.
Tech: Python, SQLite, VS Code
KCA University Intelligent Academic Advisor — A web-based academic advising system using Retrieval-Augmented Generation (RAG), semantic search, FAISS, and a locally hosted language model to answer student questions using official university documents.
Tech: Python, Flask, FAISS, Sentence Transformers, Ollama, HTML, CSS, JavaScript
The following experiences are job simulations, not employment roles.
- Customer churn analysis and predictive modeling using Logistic Regression and GridSearchCV.
- Data preprocessing, feature engineering, model evaluation, and feature analysis.
- Customer review data analysis and data visualization.
- Predictive modeling of customer buying behaviour.
- Financial analysis of Microsoft, Tesla, and Apple 10-K data using Python and pandas.
- Development and testing of a rule-based financial chatbot.
- Advanced machine learning, Generative AI, RAG, and NLP.
- Advanced Python, SQL, and data engineering.
- Business intelligence and production application development.