In Kenya, a young team of innovators is using robotics to try to break down one of the barriers facing students with hearing impairments—access to education in sign language.

The technology is still developing, but its creators say it could eventually help make technical subjects more accessible to learners with hearing loss.

In a classroom in Nairobi, communication happens largely through movement. But for deaf students, learning complex subjects can be a different kind of challenge, especially when the vocabulary needed to explain mathematics, science or technology simply does not exist in the signs they know.

"I teach physics, and in most cases, we realize that the concepts that we teach and the way I would express a concept, it becomes a challenge at some point because of the signs and the vocabulary that is needed to express a given concept," Eve Awino, a teacher at Kasarani Treeside Secondary School for the Deaf said.

That gap is what Kenyan start-up ZeroBionic is trying to solve.

At its Nairobi lab, engineers are teaching machines to communicate in sign language. For co-founder Norah Kimathi, the idea began with something much simpler.

"We started with just basic robotic hands. Since sign language is all using hands, we decided, okay, let's start with the hands and see how that goes. Then we got feedback that in sign language, there's a lot that goes into it," Kimathi said. "For example, doing this means over five different things in a certain sign language, based off of your facial expressions. That's when we decided, okay, let's build actually the full humanoid, inclusive of the face, inclusive of the eyes, inclusive of just everything to pick up on all the non-verbal cues that are currently present."

But teaching a machine to sign requires more than hardware. It needs data. When ZeroBionic started looking for African sign-language datasets, the team says there was very little to work with.

At the lab, each movement is captured, recorded and turned into data that can train the system.

ZeroBionic says the challenge now is to make the technology smarter, more accurate and more useful in the real world.

For the students watching it take shape, the technology represents something bigger than a machine.