About
Highly motivated Entry-Level Data Scientist and Machine Learning Engineer with a proven ability to develop and deploy robust AI solutions. Expertise in building predictive models, NLP applications, and computer vision systems, leveraging Python, TensorFlow, and scikit-learn. Demonstrated success in optimizing model accuracy, managing large datasets, and constructing end-to-end ML pipelines. Eager to apply data-driven problem-solving to complex real-world challenges and drive impactful business outcomes.
Work
Summary
Contributed to data science projects, focusing on data preprocessing, model development, and deployment of AI solutions.
Highlights
Cleaned and preprocessed over 10,000 data records, reducing data processing time by 30% and significantly improving pipeline efficiency.
Built and fine-tuned machine learning models for critical classification tasks, achieving 95% accuracy in heart disease prediction and 96% accuracy in diabetes prediction.
Developed and deployed interactive web applications using Streamlit and Flask, enabling real-time model inference and seamless user interaction.
Designed and implemented impactful data dashboards using Matplotlib, Seaborn, and Plotly, leading to a 30% increase in stakeholder engagement and data-driven decision making.
Publications
Skills
Programming & Tools
Python, SQL, Unix/Linux, Git.
Machine Learning
Classification, Regression, Clustering (KMeans), Association Rules (Apriori), Model Evaluation, Scikit-learn.
Deep Learning
TensorFlow, Keras, Neural Networks (ANN, CNN, RNN, DNN).
Natural Language Processing (NLP) & Large Language Models (LLMs)
NLTK, Chatbot Development, Text Classification, Summarization, Sentiment Analysis, LangChain, ChromaDB, Hugging Face Transformers.
Data Handling & Feature Engineering
Pandas, NumPy, PCA, SMOTE, Encoding.
Visualization & Deployment
Matplotlib, Seaborn, Plotly, Streamlit, Flask.
Computer Vision
OpenCV, DeepFace.
Projects
AI-Powered Customer Support Chatbot
Summary
Engineered and deployed an intelligent banking support chatbot utilizing Gemini 2.0 Flash, LangChain agents, and Google embeddings with ChromaDB for real-time document retrieval. Trained the chatbot on extensive FAQs from major Nigerian banks, automating responses to customer inquiries on loans, cards, and accounts and improving customer service efficiency. Deployed via Streamlit for seamless user interaction.
Credit Score Classification Model
Summary
Developed and deployed a multi-class classification model using scikit-learn to predict credit scores ('Standard', 'Good', 'Bad'), enhancing credit risk evaluation. Engineered features and preprocessed over 10,000 data records with pandas, achieving strong predictive performance. Deployed interactive web applications using Streamlit and Flask for real-time inference.
Text Summarization API
Summary
Developed and deployed a Flask-based REST API for automated text summarization using the facebook/bart-large-cnn model from Hugging Face Transformers. Designed to generate concise summaries from lengthy text inputs, supporting real-time applications in content curation and customer support tools.
Fraud Detection Model
Summary
Built and optimized a robust fraud detection model using Random Forest and XGBoost to accurately identify suspicious financial transactions. Effectively handled class imbalance with SMOTE and fine-tuned model performance using GridSearchCV and RandomizedSearchCV, achieving 92% accuracy and significantly improved recall through advanced feature engineering and pipeline integration. Deployed the solution via Streamlit for accessible use.