At the University of Zimbabwe, I completed a dissertation on explainable AI for forecasting next-day stock-price direction across four Zimbabwe Stock Exchange companies. My models achieved up to 76.86% directional accuracy.
I compared ARIMA, Random Forest, XGBoost, LSTM, and a Soft Voting Ensemble using walk-forward temporal validation. I also interpreted predictions with SHAP, LIME, and Integrated Gradients.
I built an interactive Streamlit dashboard for forecasts, model comparisons, and explainability. For the collaborative Crop Care Connoisseurs project, I developed an AI-powered web application to detect maize leaf diseases from images.
At Zvishavane District Hospital, I managed, cleaned, and validated routine health data using DHIS2 and Excel, and supported staff with reporting and IT issues. At Mandava Clinic, I updated patient records and supported review scheduling and prescription adherence tracking.
