I'm a third year Computer Science student at the University of Lagos, focused on building practical machine learning systems. My approach is simple: learn by doing. Every project I build teaches me something new about how to architect better models, write cleaner code, or solve harder problems.

Right now I'm getting deep into neural networks and their applications. I work across computer vision, natural language processing, and classical ML techniques, focusing on understanding how different architectures actually work under the hood. My projects range from image classifiers and regression models to distributed systems with real time intelligence. The goal is to become someone who can design and deploy production grade ML systems that solve real problems.

Outside of building models and debugging pipelines, I'm a massive Liverpool supporter. Been following the Reds for 8 years now, through the highs and lows. I also spend way too much time watching anime, especially Naruto, and binging series like Breaking Bad and Game of Thrones. There's something about great storytelling that always pulls me in, whether it's on screen or in the patterns hidden in data.

Federated Cybersecurity Platform

A decentralized phishing detection system. I engineered the Federated AI Core and Flask Backend, using LinearSVC for 91% accuracy and Redis pub/sub for network-wide threat propagation.

Deep Learning Without Frameworks

Complete implementation of a CNN using only NumPy. Built convolution layers, pooling operations, backpropagation, and forward pass from scratch to demonstrate mastery of the underlying math.

Classical ML & Feature Engineering

Waste classification system using Classical Machine Learning algorithms (SVM/Random Forest) instead of Deep Learning. Focuses on feature extraction techniques to achieve efficient classification on edge devices.

End-to-End ML Deployment

Complete pipeline from data cleaning to production. Handles outlier detection, feature engineering, and location-based price analysis. Deployed as a Flask web app with a prediction API.

Practical Data Science Guide

Comprehensive tutorial on solving class imbalance. Explores SMOTE, class weighting, ensemble methods, and evaluation metrics. Solves a common real-world industry problem.

Complex Image Classification

Optimized neural network for 10-class image recognition. Explored deep architectures and regularization techniques for high accuracy on complex datasets.

Working with convolutional networks, recurrent architectures, and transformer models. Exploring their applications in computer vision and natural language processing. Building projects that go beyond tutorials to understand how these systems actually work in production.