Dr Lighton Phiri delivered a full-day training session on Applied Artificial Intelligence and Machine Learning at the 2026 ICTAZ Tech Conference Training, EXPO and 17th Annual General Meeting, held at the Avani Resort in Livingstone from March 17 to 20, 2026. The training took place on March 17, 2026, at the invitation of the Information and Communications Technology Association of Zambia (ICTAZ), and focused on large language models (LLMs) and retrieval-augmented generation (RAG).
What the training covered
After setting up Jupyter notebooks, Google Colab and Ollama, participants worked through three parts:
- Understanding LLMs: how LLMs work, tokens and context windows, the main families of models, and how to evaluate open models on Hugging Face. The session made the case for running LLMs locally, using examples from the Zambia Revenue Authority and the National Assembly of Zambia and the requirements of sections 70 and 71 of Zambia's Data Protection Act of 2021 on transferring personal data outside the country;
- Working with Ollama: the Ollama command-line interface, Modelfiles and REST API, with a case study in which participants built a custom "Parliamentary Analyst" model; and
- Hands-on RAG: building a retrieval-augmented generation system with Open WebUI over Zambian documents.
The session also introduced the work of the DataLab Research Group, including the Enterprise Medical Imaging in Zambia project.
