Data & AI Engineer | Azure · Databricks · PySpark · LLM Agents | Ex-Mercedes-Benz MO360 Platform

Cloud-native vaccination record platform — Azure Functions, Queue triggers, Blob Storage & SQL. Runs 100% locally via Docker + Azurite. No cloud credits needed.

SchemaForge Studio | Professional database engineering showcase for Khulisa Commerce. Features 3NF schema design, advanced T-SQL (Stored Procedures, Analytics Views, Window Functions), NoSQL (Mongo…

A production-grade, full-stack analytics platform to visualize real-time, multi-city sensor telemetry. Designed to bridge the gap between high-volume backend data processing and intuitive, low-late…

A secure, full-stack government bulletin board system featuring HTTPS/TLS encryption, JWT authentication, and advanced protection against OWASP Top 10 vulnerabilities like NoSQL Injection, XSS, and…

Enterprise-grade VoC data pipeline using a Medallion Architecture (PySpark & Delta Lake) to process 1M+ Amazon reviews. Features distributed text normalization, salted joins for skew mitigation, an…

End-to-end NLP sentiment analysis pipeline for Amazon Appliance reviews. Compares LSTM, Conv1D, and BiLSTM architectures using PySpark and GloVe embeddings, featuring automated CI/CD validation.

Sequential Financial Fraud Detection using a 2-Layer GRU with Bahdanau Attention. Features PySpark sequence engineering for 590k+ transactions and ZAR business impact optimization (R1.69M+ net bene…

End-to-end fraud detection — 500k SA transactions · PySpark · XGBoost vs RF vs LR vs MLP · SHAP · MLflow · PR-AUC 0.709

Production-grade real-time ELT pipeline using PySpark Structured Streaming and Delta Lake. Replicates a high-impact architectural migration from Mercedes-Benz to achieve exactly-once upsert semanti…

End-to-end Medallion data pipeline for SA consumer debt portfolio — Azure Databricks, ADF, PySpark, ADLS Gen2

This repository contains a hands-on tutorial demonstrating the creation of ETL pipelines with World Bank data (indicators and projects). It explores various data extraction methods (CSV, JSON, APIs…

Leveraged transfer learning with a pre-trained MobileNetV2 model (TensorFlow/Keras) for image classification of cats and dogs. Implemented data preprocessing pipelines, performed feature extraction…

This project aims to provide a comprehensive overview of modern data architecture concepts, including data lakes, data meshes, cloud-based solutions, and real-time processing, and their application…

This project focuses on performing customer segmentation using Principal Component Analysis (PCA) and K-Means clustering. The goal is to understand customer behaviors based on their demographic and…

This project demonstrates the use of Scikit-learn's Pipeline for streamlining machine learning workflows. It provides practical examples using various datasets (sinusoids, cancer, Titanic, and Iris…

This project focuses on building a regression model to predict the burned area of forest fires using meteorological and spatiotemporal data. The dataset used covers forest fires in Portugal’s Monte…

Explores the design and implementation of a modern, adaptable data infrastructure using microservices.

FarmCentral: An Enterprise Resource Planning (ERP) Model Developed with ASP.NET Web Forms

The aim of this research project was to investigate the effectiveness of various machine and deep learning algorithms for trading bot development.

This project involved analyzing Amazon user reviews to determine the sentiment expressed (positive, negative, or neutral) using Recurrent Neural Networks (RNNs). The project leveraged deep learning…

This project focuses on using machine learning to detect unusual or unexpected events (anomalies) in video surveillance footage as they happen (real-time).

C# Educational Dewey Decimal System game using advanced data structures that we learnt in our final year: lists, trees and dictionaries