About Me

I hold a Bachelor's degree in Mechanical Engineering from the University of Lagos, which has provided me with a strong foundation in mathematics, analytical thinking, and problem-solving skills. My academic background complements my extensive experience in data analysis and machine learning.

During my time as a Data Analyst at Cleanpick Green Nigeria Limited, I worked extensively with data, extracting valuable insights to support decision-making processes. Additionally, I have pursued further education and certifications to enhance my expertise, including:

• Udacity AI Programming with Python Nanodegree (through the AWS AI & ML Scholarship)
• IBM Data Science Professional Certificate
• IBM Machine Learning Professional Certificate
• Google Cloud Machine Learning Professional Certification

I have developed proficiency in Python programming, and my technical toolkit includes popular frameworks and libraries such as Pandas, Seaborn, Plotly, Scikit-learn, PyTorch, and TensorFlow. Furthermore, I am skilled in SQL, Docker, and Google Cloud ML, making me adept at handling large-scale data processing and machine learning tasks.

Currently, I am actively engaged as a Freelance Data Scientist, working on diverse projects showcased on my portfolio website. My freelancing journey has included impactful projects such as:

• Predicting UK Tourism Trends: Time-series modeling to predict tourism trends using SARIMA, achieving an R² of 0.753.
• Evaluating Safety and Adversarial Robustness in Advanced Language Models (LLMs): Assessing adversarial robustness in models like Meta LLaMA 3.1 and Gemini 1.5 Flash.
• Post-Pandemic Labor Market Dynamics: Using ARIMA, SARIMA, and XGBoost to forecast unemployment trends.

Throughout my career as a Machine Learning Engineer and Data Scientist, I have been deeply involved in designing, developing, and deploying machine learning models to solve a variety of challenges. My expertise spans:

• Large Language modeling
• Forecasting and predictive modeling.
• Recommender systems and sentiment analysis.
• Time series analysis and deep learning applications.
• Data visualization, feature engineering, and model evaluation.

Additionally, I’ve contributed to developing AI-powered tools, such as the Mitan Energy AI Chatbot Application, which integrates LangChain, LangGraph, and Google Generative AI APIs to provide intelligent responses and document retrieval services.

I have received formal recognition of my skills through certifications and scholarships, which validate my mastery of machine learning techniques like supervised learning, unsupervised learning, reinforcement learning, and deep learning.

Some of my noteworthy projects include:

• Mitan Energy AI Chatbot: An interactive, AI-powered assistant that provides insights about Mitan Energy Company Ltd. It uses natural language processing (NLP) and retrieval-based models to offer accurate responses, integrated with LangGraph, Google Generative AI APIs, and LangChain, deployed on Streamlit for a user-friendly interface.
• FoodieFinders: A content-based restaurant recommender system that suggests restaurants to users based on features like cuisine type, state, and city. It uses a similarity matrix for personalized recommendations, implemented using Streamlit for an interactive experience.
• MovieLens Recommender+: A collaborative filtering-based movie recommender system using k-Nearest Neighbors (k-NN) to suggest personalized movie recommendations. The web app allows users to receive top-10 movie suggestions based on their preferences.
• NatureVision: A deep learning neural network project that classifies images of natural scenes using transfer learning, achieving an accuracy of 95.65%. The project focuses on environmental applications and allows users to upload images for classification.
• Climate TempTrend Hub: A project analyzing global temperature anomaly data with time series forecasting models like ARIMA, ETS, Prophet, GBR, and LSTM. LSTM was found to be the most accurate model, providing insights into climate trends and variability.
• Sentistranger (Sentiment Analysis): A sentiment analysis project on Twitter data related to the Netflix series "Stranger Things." It uses web scraping and the Transformers library to analyze the sentiment of tweets, providing insights through visualizations.
• Customer Segmentation: Analyzing customer segmentation using clustering techniques with K-Means, based on purchasing behavior. The project provides valuable insights for business strategies using visualizations and the Elbow Method for optimal clustering.

This information reflects my professional journey up to December 2024. You can explore these projects and more on my portfolio website: https://davidsonity.github.io.

Thank you for taking the time to learn about my experience and expertise.