I'm the Co-Founder & CEO of Addis AI, where I lead the company building voice-first AI infrastructure for African languages.
Over the last few years, our team has built speech recognition, text-to-speech, and large language models for Amharic and Afaan Oromo. Today, the infrastructure serves more than 150,000 users, supports over 1,500 developers, and has processed more than 20 million API calls. Our latest release introduced 28 production-ready voices that outperformed the major AI platforms we evaluated for these languages in naturalness and expressiveness.
At Addis AI, we've made a deliberate decision to build infrastructure.
There are too many languages, too many communities, and too many real problems for one company to solve alone. Instead of trying to build every application ourselves, we've focused on building the models, datasets, APIs, and developer tools that make it easier for other people to build. That's also why we've open sourced models and released datasets. The more people building, the stronger the ecosystem becomes.
Today, my role sits between research, engineering, product, and company building. I work closely with our teams on the technical direction of our models while helping decide where we invest our time, who we build with, and which problems are worth solving first.
The part I enjoy most is seeing developers, researchers, startups, enterprises, and governments use the infrastructure in ways we never planned for. That's one of the best parts of building infrastructure other people can use.
AI is one of the few technologies where African engineers still have the opportunity to help shape what's built next instead of always playing catch-up. I hope Addis AI makes that a little easier.
