
Streamlined laundry operations with a multi-payment kiosk solution, integrating Snapscan, Zapper, and Halo Dot Bank Payments for seamless transactions in potentially 225 laundry sites.

Enhanced factory worker experience with card-based PPE vending automation in manufacturing plants; facilitated Bluetooth communication for item selection and dispensing.

Optimized card creation process with a user-friendly, time-efficient solution; integrated specialized printer for on-demand card production.

Developed a Thingsboard-based dashboard to track solar plant performance against hourly predictions, providing data visualizations and 7-day rolling averages for 26+ plants using a Python Flask app as a proxy to process the data.

Added button-press tracking, shift-aware location monitoring, and settings verification features; expanded cross-platform compatibility. Scaled the app to support over 5000 merchandisers across 6 organizations in South Africa.

Improved UI and fixed various bugs in this React Native application used by laundry managers to track devices, earnings, and remotely activate machines.

Developed CNN for breast cancer masking level classification (Python, TensorFlow), outperforming pre-trained VGG16 model in benchmark testing.

Leveraging a heart failure prediction dataset from Kaggle, this project employs various deep learning models for binary classification to predict the presence or absence of heart disease.

Built an NLP-based fake news classifier (88% accuracy), demonstrating text analysis and evaluation skills using precision, recall, and F1-scores.

See GitHub for projects in F1 leaderboard analysis, music trends, web scraping, WhatsApp chat analysis, and more.