Building AI That Speaks Tonga and Lozi for Zambia

What happens when global AI tools meet local African languages? That is the exciting question at the heart of this project by Jeffrey Mdala, an AI Engineer | Software Developer | Telecommunications & Electronics Engineer based in Lusaka, Zambia. In this initiative, Jeffrey Mdala sets out to explore a practical and deeply meaningful challenge: whether a chatbot inspired by the capabilities people associate with ChatGPT can be built to communicate in Tonga and Lozi, two important languages spoken in Zambia.

This is more than a technical experiment. It is a step toward making artificial intelligence more accessible, more inclusive, and more relevant to everyday people across Zambia. By documenting the journey, sharing the code, and explaining the thinking behind the build, Jeffrey Mdala is doing something especially valuable for the African technology ecosystem: turning innovation into a learning process that others can follow, improve, and adapt.

That approach reflects the kind of work Jeffrey Mdala is known for through eskulu, a Zambian EdTech company building AI-powered learning platforms, and through his broader consulting and development work. It also speaks to his strength as a grounded technologist who connects advanced AI with real community needs.

Why Language Matters in African AI

Across Africa, conversations about AI often focus on the most widely supported global languages. Yet for many communities, the real promise of AI will only be unlocked when systems can understand and respond in the languages people use at home, in local markets, in classrooms, and in community settings.

In Zambia, building a chatbot that can speak Tonga and Lozi has clear social value. Language is not just a communication tool; it carries culture, identity, trust, and access. If AI systems only work well in English, then many people remain excluded from the full benefits of digital transformation. But if those systems can interact in local languages, they become more useful for education, information access, digital services, and community engagement.

This is why Jeffrey Mdala’s project stands out. It is rooted in a real Zambian context, not a generic technology trend. Instead of asking what AI can do in theory, he is asking what AI can do for people in Zambia in practice.

A Practical Experiment With Community Impact

In the original video, Jeffrey Mdala introduces a simple but powerful idea: we all know about ChatGPT, but can it speak languages like Tonga and Lozi? From there, he outlines a clear mission. He will try to figure it out, build a chatbot, document the process, upload the code, and share the ideas behind it, all while keeping the broader benefit to Zambian communities in view.

That combination of building and documenting is important. In many cases, promising technical ideas remain hidden behind private experiments. Jeffrey Mdala is taking the opposite route. By making the process visible, he creates room for collaboration, learning, and future innovation. For developers, students, researchers, and founders across Zambia and beyond, that openness can be just as valuable as the final product.

It also reflects the mindset of a capable engineer who understands that sustainable innovation is not only about shipping tools, but also about building knowledge ecosystems around them.

What It Takes to Build a Tonga and Lozi Chatbot

Although the transcript introduces the project at a high level, the technical challenge behind it is significant. Building a chatbot for underrepresented African languages requires more than plugging in a model and expecting perfect results. It often involves thoughtful work in areas such as:

• Language data collection for Tonga and Lozi where public datasets may be limited
• Prompt design and evaluation to test how well models understand local phrasing
• Translation and response quality checks to ensure useful, respectful outputs
• Documentation and code sharing so the work can be improved over time
• Community relevance so the chatbot solves real local problems rather than serving as a demo alone

These are exactly the kinds of challenges suited to someone with Jeffrey Mdala’s background. His work spans AI engineering, NLP systems, generative AI, software development, cloud solutions, and EdTech platforms. Through eskulu, he is already operating at the intersection of AI and education, making this local-language chatbot project feel like a natural extension of his mission.

His credentials also reinforce that credibility. Jeffrey Mdala’s experience includes work as an AI Engineer, including his previous role at Unicaf, and certifications such as Amazon Bedrock and AWS Lambda Foundations. Those foundations matter when building modern AI systems that may need scalable cloud architecture, experimentation pipelines, and reliable deployment strategies.

Why Documentation Is as Important as the Build

One of the most promising parts of this project is Jeffrey Mdala’s commitment to documenting the process. In African tech, documentation is often undervalued, yet it is one of the fastest ways to multiply impact. When builders explain what they tried, what worked, and what did not, they reduce the barrier for others who want to solve similar problems.

For a project like a Tonga and Lozi chatbot, documentation can help answer practical questions such as:

• How do you test AI performance in languages with fewer digital resources?
• What methods work best when training data is scarce?
• How do you evaluate usefulness from a Zambian community perspective?
• What tools and architectures are realistic for local developers and startups?

By sharing the code and ideas behind the chatbot, Jeffrey Mdala contributes not only a product concept but also a pathway for future builders. This is especially relevant for students, early-career developers, and innovation hubs across Zambia who are looking for examples grounded in African realities rather than imported assumptions.

The Bigger Opportunity for Zambia and African Innovation

This project points to a much bigger opportunity. If chatbots can support Tonga and Lozi effectively, the implications extend far beyond a single demo. Local-language AI could support:

• Education through tutoring and learning assistance in familiar languages
• Digital inclusion for users less comfortable engaging in English
• Public information access in more understandable and trusted formats
• Customer support for businesses serving multilingual communities
• Cultural preservation by bringing African languages into modern digital systems

For eskulu, this kind of work aligns naturally with the future of AI-powered learning in Zambia and across African markets. Educational technology becomes more powerful when it meets learners where they are linguistically and culturally. A student who can ask questions in a familiar language may engage more confidently, understand concepts more deeply, and feel that technology was built with them in mind.

That is why this effort matters. It is not only about whether a chatbot can respond in Tonga or Lozi. It is about whether African communities can shape the next generation of AI in ways that reflect their languages, their priorities, and their lived realities.

Jeffrey Mdala’s Role in Building Relevant African AI

There is something particularly encouraging about seeing Jeffrey Mdala lead work like this from Lusaka, Zambia. His profile combines strong engineering capability with a practical sense of purpose. With academic training in both Telecommunications & Electronics Engineering and Computer Science, and with experience across AI, software, cloud, and data science, he represents the kind of multidisciplinary talent that African innovation ecosystems need.

His recognition, including the Business With a Purpose Award at X Pitchathon in 2023, feels especially relevant here. A chatbot for Tonga and Lozi is not innovation for its own sake. It is the kind of purposeful technology project that aims to create real social value. That makes Jeffrey Mdala’s work worth following closely.

Whether through eskulu or through his consulting venture, MAY and Company, Jeffrey Mdala continues to show how African engineers can build AI solutions that are technically credible and locally meaningful at the same time.

Conclusion

At first glance, the question seems simple: can a chatbot speak Tonga and Lozi? But beneath that question is a much deeper vision for AI in Zambia. It is a vision where local languages are not left behind, where communities benefit directly from emerging technologies, and where builders openly share their process so others can learn and contribute.

By taking on this challenge, Jeffrey Mdala is highlighting an important frontier for African innovation. He is not just experimenting with a chatbot. He is helping push the conversation toward inclusive AI that reflects Zambia’s linguistic and cultural diversity. That makes this project both technically interesting and socially significant.

As this journey unfolds, it will be exciting to see what Jeffrey Mdala discovers, what the chatbot can achieve, and how the lessons from this project might influence future AI tools built in and for Africa.

Call to Action: If you are interested in AI solutions for education, local-language technology, cloud-based AI systems, or digital products tailored for African markets, keep an eye on eskulu and the work of Jeffrey Mdala in Lusaka, Zambia. For consulting collaborations in AI engineering, software development, EdTech, or digital transformation, you can reach Jeffrey Mdala at jeffmdala@gmail.com.