Organisation & Mandate
Sunbird AI is a non-profit artificial intelligence research organisation based in Uganda. Our mandate is to build practical, locally relevant, open-source AI technologies and data systems that empower public institutions, civil society, the private sector, and non-profits to drive sustainable development across Africa.
Call to Action Focus
Sunbird AI directly addresses the AI and Digital Innovations for Development priority area of the 2025 Call to Action. We focus on bridging critical evidence gaps by leveraging data and open-source technology to strengthen how evidence reaches the institutions and communities that rely on it.
Activities and Impact
Language is one of the biggest barriers to evidence reaching the people it is meant to serve. Policies, health information, and public services often fail to land because they are not delivered in the languages people actually speak. Sunbird AI's SALT (Sunbird African Language Technology) initiative was built to close that gap, developing open datasets, models, and tools for African languages historically overlooked by mainstream AI. Its flagship product, Sunflower, is a suite of open-weight large language models offering translation, summarisation, and analysis across 31 languages spoken in and around Uganda: 29 Ugandan languages plus Swahili and Kinyarwanda, extending access to the wider East African region.
We apply the same evidence-first approach to environmental AI, where the gap is physical as well as informational. Sunbird AI designs and builds custom embedded hardware, including low-cost, autonomous sensors that classify sound at the edge without relying on constant connectivity or cloud processing. Working with the Kampala Capital City Authority and Entebbe City Authority, we deployed these sensors to map the acoustic landscape of Uganda's cities, producing the Urban Noise Uganda dataset, one of the largest urban noise datasets in Africa.
We extend this to energy access as well. In Lamwo District, Northern Uganda, Sunbird AI used satellite imagery and open geospatial datasets to build a site identification tool that recommends where new electricity connections will have the greatest impact. Developed with GIZ and Uganda's Ministry of Energy and Mineral Development, the tool supports the National Electrification Strategy's goal of connecting 10 million households by 2030.
Staff Quote
"At Sunbird AI, we build practical AI tools and technology that reflect the languages, environments, and realities of the communities we serve. This work equips institutions with the evidence they need to plan effectively and govern with confidence."
— Dr. Ernest Mwebaze, Executive Director, Sunbird AI
Mr. Muganzi, a coffee farmer in the Mityana District in Uganda, uses Sunflower to translate crop conditions and farm management.
Permanent Secretary, Ministry of ICT, Uganda, Dr.Aminah Zawedde, launches Sunflower, a large language model built by Sunbird AI.