ABOUT STHEMBISO MKHWANAZI
Sthembiso Mkhwanazi is a purpose-driven researcher and engineer specialising in Natural Language Processing (NLP) and machine learning for African languages. He holds a BSc in Computer Science and Data Science and a BSc Honours from the University of KwaZulu-Natal (UKZN), and recently completed an MSc in Computer Science at North-West University (NWU) , focusing on the morphological segmentation of isiZulu.
Sthembiso most recently served as a researcher in the NLP Research Group at the Council for Scientific and Industrial Research (CSIR), where he built and deployed language technologies for isiZulu, Sepedi, and Afrikaans, including automatic speech recognition systems, a children's reading assessment app piloted in Gauteng schools, and an AI-powered story generation platform . He has since joined Discovery Limited as a Machine Learning Engineer, bringing his NLP and AI expertise to one of South Africa's largest financial services and technology companies, which operates across 42 markets globally.
He has published research on recommendation systems and retrieval-augmented story generation, machine translations, etc., and has reviewed papers for international venues including WMT, and serves on the organising committee for IndabaX South Africa 2025 and 2026. His acceptance into Black in AI and the Fatima Fellowship reflects his commitment to advancing globally relevant African AI research. He has also spoken on various radio stations, communicating the value of AI and language technology to millions of listeners in isiZulu and siSwati.
Alongside his research and industry career, Sthembiso lectures at Eduvos, teaching modules in Artificial Intelligence, Machine Learning, Mobile Development, and Game Design. He is committed to mentorship and youth development, advocating for greater STEM participation from disadvantaged communities. He aspires to pursue a PhD in NLP focusing on African languages, multilingual learning, and responsible AI.
The work of Sthembiso Mkhwanazi
Sthembiso Mkhwanazi has been actively involved in various projects within NLP, a subfield of Artificial Intelligence. His journey in this field was ignited during his Honours mini-thesis, where he developed a news recommendation system that leveraged social media signals to deliver personalised content. This project not only demonstrated strong technical and research capability but also earned him the Best Research Award in 2022. The experience sparked a deeper curiosity about intelligent systems and their real-world applications.
Following this achievement, he joined the CSIR as a Junior Researcher in NLP. At the CSIR, he has contributed to the development and training of machine learning systems focused on low-resource African languages. His work has included building and improving machine translation systems, training Automatic Speech Recognition (ASR) models, and implementing Retrieval-Augmented Generation (RAG) approaches for isiZulu story generation aimed at addressing early childhood literacy challenges.
Through these initiatives, he has contributed to research outputs published in reputable journals and conference proceedings. He has also been involved in strategic projects such as Ngiyaqonda at the CSIR, which seeks to support foundational literacy development through technology-driven interventions, and spoke in various seminars, retreat, conferences including several radio stations. His academic excellence is reflected in graduating with cum laude and distinction in his degrees, underscoring both his technical competence and research dedication.
[1] https://www.youtube.com/watch?v=53qiaaSBiJM&t=14s [2] https://link.springer.com/chapter/10.1007/978-3-031-78946-5_7 [3] https://www.kaggle.com/code/spikemayor/tweets-classification-rnn [4] https://scholar.google.com/citations?user=5aHvh04AAAAJ&hl=en [5] https://m.youtube.com/watch?v=gmQiqx8x0J8 [6] https://www.facebook.com/PanSALB/videos/to-help-mitigate-the-literacy-levels-in-south-africa-the-csir-developed-the-ngiy/2298158093989883/ [7] https://www.dsti.gov.za/index.php/media-room/latest-news/4504-ngiyaqonda-literacy-app-to-boost-home-language-education-in-south-africa [8] https://journals.co.za/doi/10.10520/ejc-csir_sci_v23_n1_a8 [9] https://scholar.google.com/citations?user=48TjKAgAAAAJ&hl=en