In a small lab at Sokoine University of Agriculture in Morogoro, Tanzania, artificial intelligence is doing something unexpected—it’s not just solving problems, it’s changing lives. The platform behind this quiet disruption? Zindi is a Pan-African data science competition hub quietly training a new class of problem-solvers on the continent. One of its most compelling champions is Catherine Francis Mangare, a 35-year-old assistant lecturer turned AI mentor, now pursuing her PhD in Japan.
Her journey from dreaming of piloting aircraft to piloting deep neural networks reveals a more profound truth: Africa’s tech transformation is being driven from the ground up, by people who were never supposed to be in the room.
The Pivot That Sparked a Movement
Mangare’s path to AI wasn’t linear. “When I was young, I wanted to be a pilot,” she recalls. But life had other plans. Poor results in geography derailed her aviation ambitions. She pivoted into ICT on a friend’s advice and never looked back. What began as a compromise soon became a calling.
She now teaches computer science and informatics while researching data mining techniques for agriculture. Her real impact, however, lies in how she uses her platform to bring others along, especially young students with zero exposure to AI. “There’s no way you can run away from artificial intelligence,” she says, “so why not run toward it?”
Zindi: More Than Just a Competition
For Mangare, Zindi wasn’t just another online challenge—it was the gateway to real-world data science. Introduced through her university’s Innovation and Entrepreneurship Lab (IES Lab), Zindi offered her a direct line to high-stakes machine learning problems beyond the classroom.
In one challenge, she helped develop models to predict waterborne diseases, first using classical ML models, then pivoting to deep neural networks when results fell short. “I saw how model performance changed, and I learned how to structure a real AI workflow—from cleaning data to building production-ready models,” she says.
But Zindi’s true power lies in accessibility. With a Python notebook, a dataset, and a timer, students are challenged to think: Can they really solve global-scale problems? Mangare used that momentum to revamp student learning at her university, adding AI content to the curriculum, creating tutorials on YouTube, and offering peer-led boot camps.
From Beans to Bots: The AI-Agriculture Frontier
If you want AI in action in Tanzania, look to the fields.
Mangare’s research focuses on disease prediction in crops like maize and beans, combining plant biology with AI image classification. Her team collaborates closely with agricultural scientists to label datasets and validate model accuracy. “It’s hyperlocal work,” she explains. “These are real diseases affecting real farmers. The impact is immediate.”
And the ambition doesn’t stop at prediction. Catherine’s mentor, Dr. Kadege Fue, built a robotic vehicle that autonomously distributes fertilizer based on sensors’ detected soil moisture, essentially precision farming, designed in a Tanzanian lab.
Zindi is filling the gap and enticing African problem solvers. Many of the ongoing issues proposed on the platform mirror real-world use cases from agriculture, climate, road traffic, service client efficiency, etc. “What we test on Zindi, we bring to the farm,” says Mangare. “It’s not theory—it’s field-ready.”
The Swahili AI Revolution
One of the biggest hurdles facing AI in Africa is language. While global LLMs have expanded support for non-English languages, tools like ChatGPT are only just beginning to integrate Swahili. But Tanzanian developers aren’t waiting around.
Mangare and her colleagues are building Swahili chatbots and NLP models from the ground up. One project, Mkulima GPT, helps farmers in rural areas get answers to crop questions in their native language. Another effort involves Swahili news classification, with datasets partially sourced from Zindi competitions.
The goal is to create a Swahili-centered AI ecosystem that doesn’t just translate global tech, but originates local innovation.
Simulations, Not Just Seminars
Even as she studies in Japan, Mangare is planning her return with clear intent. “We can’t always afford expensive machines, but we can simulate them,” she explains. Her next focus is building simulation-based curricula that let students design AI systems in virtual environments—robotics, precision agriculture, medical diagnostics—without the high price of hardware.
And she’s not alone. Many of her former students have gone on to build AI tools post-graduation, from diet-planning apps to disease-prediction models, often in partnership with mentors from the Zindi community.
Scaling AI From the South
Zindi may not have the brand cachet of Kaggle or OpenAI, but its impact across Africa is tangible. It’s a platform that democratizes access to cutting-edge problems—water sanitation, climate data, public health—and it does so in a way that prioritizes African languages, infrastructures, and realities.
Mangare’s story is a microcosm of this movement. A single lecturer, motivated by a missed childhood dream, inspires dozens of young Tanzanians to become leaders in AI. Some might end up building the next big LLM. Others might design a robot that saves harvests. All of them are proof that talent is everywhere—opportunity isn’t.
That’s where Zindi comes in. It’s not flashy. It’s not VC-funded hype. It’s just working.
//Staff writer
