Ndzhaka

Ndzhaka was an information knowledge hub built for the National Treasury in South Africa. The project involved creating a system to crawl both public and private government websites, extract documents and information, and load them into Azure Cognitive Search for intelligent indexing and retrieval. I played a key role in developing the crawler using Python, building the backend with .NET Core, and implementing the search functionality. I also contributed to the Angular frontend and managed the CI/CD pipelines. A significant challenge was optimising search relevance and accuracy, which we addressed by fine-tuning the search algorithm and carefully sanitising the data. The result was a powerful search tool that allowed National Treasury employees to easily find documents and resources across all government domains.