Michael Chinaka

Recent university graduate with a proven track record of developing high-quality software solutions. Committed

to continuous learning and staying up-to-date with the latest technological developments. Collaborative team

player with strong analytical and problem-solving skills, ready to contribute to a graduate level software

engineering role.

Premier League Results Predictor Dissertation

• Developed a predictive model to forecast English Premier League match outcomes using Python and machine learning techniques.
• Conducted data collection by web scraping with BeautifulSoup and performed data pre-processing using Pandas in Jupyter Notebooks.
• Utilised Scikit-learn for comparative analyses of various machine learning techniques, creating an optimal predictive algorithm.
• Implemented comprehensive data pre-processing, feature selection, and hyperparameter tuning to optimise model performance.
• Designed and deployed a user-friendly website with Streamlit, hosted on an AWS EC2 instance, enabling user interaction with the model and receiving predictions.

Simulation Framework (IBM)

• Automated process of pre-analysing data file, ingesting it into a database and training the data using our AI algorithms with the use of bash scripts.
• Created a Java tool to analyse the results obtained from training to provide in-depth analysis.
• Utilised object-oriented inheritance to create a hierarchical, tree-like data structure representation. Performed comprehensive analysis of the structured data.
• Implemented error handling, exception management, and comprehensive unit tests using JUnit for reliability and correctness.

Artificial Neural Network for Predicting Mean Daily Flow

• Developed a multi-layer perceptron (MLP) model in Java to predict river flow at Skelton using upstream flow and rainfall data.
• Processed and cleaned the dataset in Excel, enhancing data quality by removing outliers and imputing missing values.
• Optimised the MLP training with advanced techniques - Backpropagation for effective error minimisation, Momentum for stabilising and accelerating convergence, Bold driver method to adaptively adjust the learning rate.

• The Project Management Course, Udemy
• Agile Crash Course: Agile Project Management, Udemy
• Enterprise Design Thinking Practitioner, IBM Python for Data Science, Cognitive Class
• Docker Essentials, IBM

• English (Fluent)
• Shona (Fluent)
• Afrikaan (Basic)