Skip to the content.

Data Scientist

Technical Skills: Python, R, SQL

Education

Work Experience

Data Science Consultant @ Amdari (February 2025 - Present)

Data Scientist @ Nigerian Ports Authority(August 2022 - January 2025)

Projects

Risk Classification Thesis Project

Thesis

Developed a script using Python to fetch section 1.A (Risk Factors) of the 10-K filings from the EDGAR Database. Researched several NLP manipulation techniques to clssify the risk factors including Linear SVC, Naive Bayes, Logistic Regression, Topic Modelling, Neural Networks and Neural Topic Modelling. With the rise of LLMs (Large Language Models), that was the best solution to the risk classification problem but it costs more. AI Solution

AI Risk Classification

Repository

Built an automated pipeline to extract risk factors from U.S. SEC 10-K filings and classify them with language models into risk categories. Ingests text from filings, preprocesses risk text, and applies NLP/LLM classification to assign risk category labels. Technologies: Python, Hugging Face/LLM API (or similar), NLP preprocessing (tokenization, vectorization), structured output classes for risk taxonomy.

FNOL(First Notice Of Loss) Intelligence System

Repository

Developed a machine learning model to predict claim outcomes/costs based on First Notice of Loss (FNOL) insurance data. Workflow included cleaning and transforming structured incident/claim features, exploratory data analysis, and supervised ML training with evaluation metrics. Techniques likely used: Python, pandas/NumPy, Scikit-learn models (e.g., regression/classification), feature engineering, model validation.

Customer Segmentation using RFM Analysis

Repository

Performed customer segmentation of banking customers using RFM (Recency, Frequency, Monetary) analysis to score and cluster clients by value/behavior. Computed RFM metrics from transactional data, then applied clustering (e.g., K-Means) to derive segments for targeted marketing. Technologies: Python, pandas for data manipulation, Scikit-learn for clustering, visualization libraries for segment profiling.

AI Bed Prediction

Repository

Developed a machine learning system to predict hospital bed demand using historical hospital utilization data, enabling proactive capacity planning and operational decision-making. Built a full data science workflow including data preprocessing, feature engineering, and predictive modeling to forecast future bed requirements. Technologies and techniques: Python, Jupyter Notebook, pandas and NumPy for data manipulation, exploratory data analysis, supervised machine learning models for demand prediction, and performance evaluation using regression metrics.

Travel Assistant AI

Repository

Built an AI-powered conversational chatbot designed to handle hotel-related queries, enabling users to interact naturally for information retrieval and basic booking assistance. The system processes user input, interprets intent, and generates context-aware responses to simulate a hotel assistant experience. Technologies and techniques: Python, natural language processing, large language models (LLMs) for conversational response generation, prompt engineering, and integration frameworks such as LangChain and API-based LLM services for dialogue management and response orchestration.

Shift Perfomance Optimisation

Repository

Developed a data analysis and predictive modelling solution to evaluate and improve workforce shift performance using operational data. The project processes shift-level data to identify patterns in productivity, inefficiencies, and performance variability, then applies analytical and machine learning techniques to generate insights and performance predictions. Technologies and techniques: Python, pandas and NumPy for data preprocessing, exploratory data analysis (EDA) for performance trend identification, feature engineering on time/shift-based variables, and supervised machine learning models (via scikit-learn) to assess and predict performance outcomes.