As part of the Edge 50 list, Apolitical is publishing a series of articles exploring a selection of these bold initiatives to learn more about the stories and strategies behind them.

Created by Apolitical in partnership with the Mohammed Bin Rashid Centre for Government Innovation, this list is an invitation to look beyond the usual playbook and ask what becomes possible when governments are willing to think boldly about today's biggest problems.

Apolitical spoke with Celina Lee, Founder of Zindi, an Africa-based platform that runs open data science competitions tied to real-world public-interest challenges, connecting over 90,000 AI practitioners across more than 185 countries with the governments, NGOs and companies that need solutions.

In much of Africa, breaking into an AI career has meant having the right degree, the right connections, or being in the right location, leading companies to assume they had to look outside the continent entirely. Zindi starts from a different premise: AI talent exists across Africa, and what's missing is the opportunity to prove it. Unlike the usual route of training people and hoping jobs follow, Zindi treats public-interest challenges as a proving ground where skills, visibility and work emerge together. Research from a partnership with the Kenyan Presidency found that nearly 1 in 5 users went on to advance their careers.

Tell us about when Zindi started. What did you see about youth unemployment and AI talent that others were missing?

We're the largest platform for data scientists and AI developers focused on Africa. We launched Zindi in 2018, and the starting point was that I had moved to South Africa for a different job with an international NGO. Part of that job was to develop local talent in Africa to apply machine learning, AI and data science techniques to development areas, especially financial inclusion.

Through that job, I met tons of young people across Rwanda, Kenya, and Nigeria who are super talented but didn't have opportunities to build their skills, showcase themselves, or connect with job opportunities. My co-founders were running a data science consulting company working with some of the biggest companies in Africa, and they were finding that companies assumed they had to look outside their country and even outside the continent to get the talent they needed.

So for us, the idea was to create a space where young people could build their profiles and skills, and where companies could find these talented people and get the solutions they needed.

Why did you decide to go down the route of real-world competitions?

For two reasons: One is that training isn't enough. We've seen young people who get degrees in computer science or mathematics, but they're not necessarily able to jump right into their first job. The most critical bridge is practical problem-solving experience that companies demand, even at a junior level. The other reason is that when we started in 2018, there were already many courses available through platforms like Coursera and Udemy, and companies like Microsoft and IBM were offering free courses. It was a crowded space. We didn't want to reinvent the wheel, but to complement what was already out there. So competitions allow people to apply their skills at scale, which is something many organisations struggle to do.

Did you ever think 'this might not actually work'?

The hardest thing is that we're trying to solve many problems at once. It's a complex problem, building AI capability in markets that haven't been part of mainstream technology spaces.

"We're building skills, creating pathways, solving real AI problems, and bringing along the private and public sector as adopters."

There's also the challenge that if opportunities don't exist locally, talent will leave. So we have this multi-pronged problem. In the early days, and even still, it was hard to reconcile all the problems we were trying to solve with one single solution, which was the Zindi platform. But we're getting closer.

Is there anything that you've decided to stop doing?

There are things we've put on pause. We're a startup and haven't raised a ton of capital, so there have been moments where we had to make very hard decisions. One lesson learned is the power of focus and saying no. Because we make the platform free and open, our users are not paying us. During periods of financial constraint, we had to pause activities that benefited users but didn't generate revenue in the short term. That forced us to prioritise the side of the ecosystem that represents revenue, making sure our challenges delivered tangible value to organisations hosting them.

How do the partnerships with competition hosts work?

For example, we're working with the government of Barbados on a longer-term partnership. It's about building the ecosystem, engaging local talent through workshops and hackathons, and raising capability. We then worked with different government agencies to select AI use cases. One competition automates the processing of hand-drawn land maps. Another uses traffic camera data to identify driving behaviours linked to congestion. A third focuses on digitising historical documents with difficult handwriting. Some partners come with clearly defined problems, like the government of Togo, and others co-create use cases with us.

What does an individual's journey on the platform look like?

It's free and open to everyone, and creating an account takes about 30 seconds. Many people join for competitions. We provide datasets and starter notebooks to help them begin. They download data, train an AI model, and submit results. The leaderboard is live, so there's an iterative, gamified process where people move up and down based on performance. It becomes very gamified. Users collaborate on discussion forums, form teams, and at the end, the top three win cash prizes. Everything contributes to their Zindi profile, which becomes a portfolio or live CV of proven skills.

What have you learned about working with governments?

Timing is critical. Governments need to have already decided that AI and digital transformation are priorities, with political will and funding behind them. We fit well when an AI strategy has just been put in place, because most strategies include building local talent and local innovation. What we do is very tangible. Within less than 12 months, we can deliver impact in terms of skills and AI use cases.

What advice would you give to public servants?

I would challenge public servants to think creatively and out of the box.

"AI skilling isn't just about courses and boot camps. Be open to alternate training models that can deliver greater scale and impact, and remember the goal is not training for its own sake. It's about connecting the journey from learning to real opportunity."

Platforms like Zindi help bring that together and help identify real problems where AI can create value.

What's next for Zindi?

We're excited about the Barbados partnership, which is the most comprehensive programme we've run, and we're looking to replicate that in other countries. The programme was a 12-month initiative focused on building and growing the Barbados AI ecosystem through capacity building, community engagement, local challenge development and talent pipeline creation. The programme onboarded students and professionals into the Zindi community. We hosted hackathons and upskilling events with local university groups and launched AI competitions based on locally relevant public-sector problem statements.

We identified practical public-sector AI use cases that could create immediate value, including automating the processing and review of land plans and applying computer vision and predictive modelling to better understand traffic patterns and driver behaviour to help reduce congestion and accidents. These competitions are designed to generate practical AI solutions while also identifying and nurturing local talent.

The programme identified top-performing participants and connected them to employment and collaboration opportunities. By linking skills development with innovation and job creation, the initiative contributed to economic growth and supported the development of a sustainable AI ecosystem in Barbados.

It's important to create quick wins and bring people along the journey. If you try to build something massive over five years, you risk losing people along the way.