I believe that behind every dataset lies a story waiting to be uncovered. I blend curiosity with rigor — exploring messy data, revealing hidden patterns, and transforming raw information into insights that drive smarter decisions. My work spans predictive modeling, natural language processing (NLP), data cleaning, exploratory analysis, and dashboarding. Whether I’m building machine learning models, breaking down business data, or analyzing user behavior, my goal is simple: turn complex information into clear, actionable insights. I’m also skilled in data annotation, including labeling images and videos, drawing bounding boxes, performing segmentation, and preparing high-quality training data for computer vision and AI systems. This hands-on experience gives me a deep understanding of how models learn and succeed in real-world environments. What excites me most is AI — not as a buzzword, but as a powerful tool. I enjoy building systems that learn, adapt, and assist; improving prediction accuracy; and solving challenging problems that push innovation forward. If you want to make data count, bring clarity to the numbers, or build intelligent, AI-driven solutions, I’d love to connect. Let’s create something truly insightful together.
Download CVI’m skilled in Python, SQL, Power BI, and Excel, which I use to clean data, analyze it, and turn it into simple, useful insights. I create dashboards, automate tasks, and help people understand what the data is saying. I’m also good at problem-solving, teamwork, communication, and staying adaptable — skills that help me work well in any data environment.
Read more
This project uses the K-Nearest Neighbors (KNN) algorithm to predict whether a breast tumor is benign or malignant based on medical diagnostic features. The model was built using Python, trained on a clean and structured dataset, and evaluated for accuracy and reliability. The final solution provides fast, interpretable predictions that support early detection and medical decision-making.
I developed a computer vision system to detect Kenyan vehicle license plates using the YOLOv8 model. The project involved collecting and annotating 300 diverse vehicle images with CVAT, training YOLOv8m for 50 epochs, and achieving a mean Average Precision (mAP) of 0.85. This solution can be applied to automated license plate recognition, toll management, parking control, and smart traffic enforcement, demonstrating practical AI applications for Kenya’s intelligent infrastructure.