How a 17-Year-Old Nigerian Built a Hybrid AI System, Attracts 4,000 Developers

A 17-year-old Nigerian self-taught developer, Okechukwu Nwaozor, has built a hybrid artificial intelligence (AI) system that is already attracting thousands of users and developers across the country.

Nwaozor, who recently completed secondary school, is the founder of OkeyMeta, an AI startup he registered after spending three years experimenting with data collection, model training, and system deployment. Despite his age, he has raised about ₦2.7 million to fund the project, which he describes as the result of relentless curiosity and self-learning.

Speaking with Techpoint Africa, Nwaozor said he was often dismissed whenever he talked about building his own AI product. “People told me it couldn’t be done,” he said, recalling the mockery that followed his early posts on Facebook announcing his ambition.

OkeyMeta is built as a hybrid AI model, combining fine-tuned open-source systems such as Google’s Gemma-2B with African-focused datasets that Nwaozor gathered and cleaned himself. He also developed custom reasoning layers, knowledge-distillation pipelines, and API infrastructure to enable developers to build on the platform.

According to him, inspiration came long before the rise of ChatGPT. His interest in AI began with Google search technology and deepened after seeing developers create conversational AI bots in online communities. By age 14, he had started building experimental models before finally settling on OkeyMeta as his flagship project.

Nwaozor currently leads a small team of undergraduate collaborators who assist with engineering and deployment. The company’s chatbot, OkeyAI, has about 900 registered users, while its APIs serve over 4,000 active developers. He said visibility remains low despite the adoption. “Our posts only get one or two likes,” he noted, adding that every feature is currently free as part of his strategy to drive awareness.

He revealed that the system includes agent-based functionalities that allow the model to take actions only under specific conditions, such as replying to an email after confirming a user’s interest in a product.

However, Nwaozor acknowledged that growth comes with high operational costs. Running the platform costs about $100 per month, a figure expected to rise as more users sign up. He said his current funding will soon run out and that some investor offers have included terms that could harm the company in the future.