Happy Herman and Diarra Niang are making artificial intelligence reachable through the one thing millions of people already have. A phone call. This is the story of **Bakame AI.**
Published: March 31, 2026
I found them through a post from the World Government Summit in Dubai. African students honored for building an AI education system that works without internet. I saw it and felt that kind of pride that is hard to explain unless you are Pan-African and you know what it means to see your people doing something the world did not expect. I looked closer and realized I already knew one of them. Happy Herman was already in my Instagram following. We had connected before I even knew what he was building.
I knew this was my next entry on Entry One.
The moment we got on the call he was already talking, already excited, already giving. That is just who he is. Happy Herman is one of those people who makes you feel like the conversation you are having is the most important one he has had all week, and he probably means it. He is smart, deeply people oriented, and curious about technology.
He grew up in Rwanda, and his dad was a high school literature teacher. Shakespeare was everywhere in the house.
When you grow up in a household of educators, the value you place on the promise of education is astronomical!
He did not have a computer growing up. Not until after COVID. It was through Bridge2Rwanda/Isomo that he got his first one, and that computer changed everything. It led to an application, an acceptance, and eventually a full scholarship to TCU in Fort Worth, Texas. He arrived in the US in fall 2024. AI was everywhere. The hype was high. Everyone was talking about ChatGPT. He used it to write an essay and immediately thought about home.
"I was like, hell no. This thing writes. I've been learning how to do this for 10, 15 years of my life. It writes it in like 30 seconds. But my people don't have access to this.
That question, how do we bring this home, became the square one for Bakame AI. He started asking professors about it. He learned about natural language processing and that it was expensive and complicated. He kept going anyway. Then he walked up to a professor, told him he wanted to change the world, and asked for $10,000. That money funded a year of building, breaking things, rebuilding. He did not have the technical knowledge yet. He was pitching to engineers before he had anything to offer them. Working on social capital and belief alone. Ninety percent of people said no. He kept going because he knew the ten percent was real. By 2025 the idea had found its shape. Deploy the AI on telephone infrastructure instead of the internet. You call a number. The AI answers. No app, no data plan, no Wi-Fi. Just a phone call, the same way you have always made one.
You may be asking yourself just like did (if you were kid in Rwanda) Bakame? wait is that the character from the books we always had as kids? lol
When people outside Rwanda see the name something clicks anyway. Education. Knowledge. Making people smarter. Happy did not plan that. It just fit.
Bakame, the clever, tenacious, self-taught figure from the books. Happy grew up loving that character. The resilience. The curiosity. The refusal to be stopped. He could not change his own name. But he could build something that carried that spirit.
I grew up feeling in love with how smart Bakame was. The tenacity, the resilience. I always wanted to be that Bakame.
So how does Bakame Ai work for anyone without a technical background? Here is the simplest way to understand it. When you call your aunt in Kigali, your voice travels through MTN's towers, hits a data center, gets processed, and comes back to her phone. No internet browser. No data plan. That infrastructure already exists everywhere across the continent and it already works. Bakame Ai runs on that same path. You call a number. The AI answers. It listens, understands your question, and responds through the same phone call. Like calling a teacher who is always available, always patient, and never runs out of time for you.
It is not trying to do everything. Happy is deliberate about this. A large language model is like a professor with PhDs in everything, impressive but expensive and heavy. It is a small language model, scoped specifically to education. Right now that means English practice, homework help across core subjects, and debate coaching through a live pilot with I-Debate and Isomo soon. Rwanda's data sovereignty laws require all data to be processed on the ground, not exported to foreign servers. That constraint forced a smarter, leaner architecture. Everything runs locally. What could have been a limitation became a feature. The technical side is genuinely hard. Latency matters in a voice call. Too much delay and the person hangs up. The system has to handle speech to text, language processing, and text to speech while staying fast, on inconsistent connections, through ordinary phone lines. Every decision has had to be intentional. Language goes beyond just making it work. Bakame Ai is being trained to speak Kinyarwanda, Wolof, and Pulaar. The goal is not just translation. It is making the AI feel natural to the person using it, in their language, in their accent, in the way they actually speak.
Happy talks a lot about how he builds his team. He goes fishing, he says, for people who are smarter than him. If you are in a room and you are the smartest person, you are in the wrong room. That philosophy led him to Diarra Niang.
Diarra is Senegalese. She grew up on the continent, studied at the African Leadership Academy in South Africa, and is now pursuing a dual degree, a BBA in Business Information Systems and a BSc in Economics with a FinTech certificate. She understands the African landscape not as an observer but as someone who has lived across it. She understands markets, finance, and what it takes to build institutional relevant solutions like Bakame.