Chukwuemeka Ezumezu

CHUKWUEMEKA EZUMEZU -, -- linkedin.com/in/chukwuemeka-ezumezu github.com/EmekaEzumezu SUMMARY Lead Machine Learning Engineer with over 5 years of experience. Results-driven and committed to leveraging data-driven insights to optimize business strategies. Strong background in computer science, skilled in cloud computing, data science, MLOps, LLM, and algorithm development. Experienced in developing and deploying machine learning models, data analysis, and software development. WORK EXPERIENCE CHEKKIT TECHNOLOGIES CORP. Lead Machine Learning Engineer ● ● ● ● ● ● ● Ikoyi, Lagos 09/2022 – Present ChekIntel - a real-time business intelligence dashboard web app product: Developed and deployed ChekIntel web app (Flask, Python Plotly, scikit-learn, Docker Container, Microsoft Azure, App Service, CI/CD pipelines with GitHub Action, ML classification model, etc.), delivering the following key features: 1. Seamless retrieval and comprehensive analysis of client data from a MySQL database, encompassing univariate, bivariate, and multivariate analysis. The dashboard displays charts and tables on the user interface, enabling companies/clients to gain real-time insights from their first-party customer/user data, and facilitating informed data-driven decisions. 2. Enhanced flexibility for clients to upload survey or sales data (CSV or pdf format), automating the end-to-end analysis, including time series analysis for future sales predictions. 3. Empowerment for clients to effortlessly generate and download comprehensive PDF reports, containing profound insights and implications extracted from the analysis. 4. These initiatives led to a significant 10% profit improvement. ChekInspect - Computer Vision Product Line Inspection: Research and development of ChekInspect, a cutting-edge computer vision solution designed to enhance product inspection processes within a production line environment. 1. Designing a robust computer vision model with the capability to precisely identify and categorically count products as they traverse the conveyor belt. Leveraging advanced camera systems and meticulously crafted ML algorithms. 2. Innovating defect detection: Creating an AI-powered system to scrutinize product sticker labels for anomalies or defects. Utilizing sophisticated cameras and machine learning models to segregate substandard products from quality ones efficiently. Conducted research and development for ChekInspect, a mobile-based product validation system using neural networks and AI for counterfeit detection through image analysis, achieving improved product authenticity and quality assurance. Developed and deployed a machine learning model to categorize user survey data based on behavioral analysis, generating actionable insights via an interactive dashboard like ChekIntel, resulting in a 30% boost in data-driven decision-making. Utilized clustering techniques to group similar data points and assign labels for improved data organization and analysis. Collected data from various sources, including web, Excel, and CSV files, and conducted data preprocessing, feature extraction, modeling, and visualization using an interactive web app. Utilized Tableau to create dynamic and visually engaging data dashboards for on-demand data presentation. PUBLIC COMPLAINTS COMMISSION (THE NIGERIA OMBUDSMAN) Data Analyst ● ● ● Lagos Island, Lagos 09/2013 – 08/2022 Conducted research, analyzed data using statistical techniques, identified trends, prepared reports, and made recommendations for the commission. Leveraged data-driven decisions to increase the closure rate of complaints by 30%. Initiated a new filing system for proactive cases, enhancing case file management and improving service delivery efficiency by 28%. SKILLS - MLOps: FastAPI, Modelling, Deployment and Monitoring, CI/CD pipelines with GitHub Action, Docker Container, App service, Synthetic Data Generation, TensorFlow Extended (TFX), Responsible AI. - Cloud and Web: GCP (AutoML, BigQuery, Vertex AI, Gen AI), AWS (Sagemaker), Azure, Flask, Streamlit. - Data Scientist: Python, NumPy, Pandas, Scikit-learn, Documentation, Communicating to Stakeholders, ETL, NLP and Machine Learning Pipelines, Experimental Design, Recommendation Systems, Linear Algebra. - Machine Learning Engineer: Scikit-learn, Supervised Learning Algorithms, Clustering, Dimensionality Reduction (PCA), Data Processing, Feature Extraction, Data visualization, Modelling and Deployment. - Data Analyst: Descriptive and Inferential Statistics, Data Wrangling, Exploratory Data Analysis, Data Visualization, A/B Testing, Matplotlib, Seaborn, PyPlot, Git and GitHub, BitBucket, Team Collaboration. - Data Structures and Algorithms: Arrays and Linked Lists, Stacks and Queues, Recursion, Tree, Maps and Hashing, Sorting, Divide and Conquer, Greedy, Graph Algorithms. - Business Analytics and Predictive Analytics for Business: Excel, SQL, Tableau, Alteryx Analytics Software. - Computer Vision: PyTorch, TensorFlow, Keras, C++, CNN, RNN, LSTMs, GAN, Intel OpenVINO™ toolkit, MQTT, OpenCV, Calculus. - Natural Language Processing: Feature extraction and embeddings, Deep Learning Attention, Topic Modeling, Sentiment Analysis. - Generative AI with Large Language Models (LLM): Prompt Engineering, Attention Mechanism, Parameter Efficient Fine-Tuning (PEFT). - Others: C, PHP, Java, Firebase, Mobile App Development (Flutter or Java), Flask Full-Stack Development, Deep Reinforcement Learning. EDUCATION - Bachelor of Science in Computer Science, National Open University of Nigeria (NOUN), January 2014 - July 2019 CERTIFICATIONS - AI Product Manager Nanodegree, Udacity, March 2021 - August 2021 - Intel® Edge AI for IoT Developers Nanodegree, Udacity, March 2020 - July 2020 - Deep Learning Nanodegree, Udacity, March 2020 - July 2020 - Machine Learning Engineer Nanodegree, Udacity, November 2019 - February 2020 VOLUNTEER EXPERIENCE OMDENA Machine Learning Engineer ● ● 07/2020 – 10/2020 Contributed to a project focused on preventing malaria infections through topography and satellite image analysis. Developed an algorithm that combines satellite images, topography data, population density, and other data sources to identify stagnant water bodies for targeted treatment, thus preventing mosquito breeding.