Digital Umuganda

" Language is more than just words; it's identity, culture, and heritage " – Digital Umuganda

Digital Umuganda is a Rwanda-based artificial intelligence and open-data company that develops language technologies for African languages. Its work focuses on reducing the digital divide by building the language infrastructure needed to ensure artificial intelligence is developed for Africans. Digital Umuganda creates high-quality datasets, develops language models and open-source libraries, and builds practical AI solutions, focusing particularly on education, healthcare, Government services and agriculture. By combining community-driven large-scale data collection, applied research, and product development, the company makes digital services more accessible to people who primarily communicate in African languages. Its broader mission is to ensure that African languages, knowledge, cultures, and communities are meaningfully represented in the development and use of artificial intelligence. To date, Digital Umuganda has collected data in more than 25 languages across 20 African countries and has made approximately 15,000 hours of open source language data available to support and accelerate language-based AI innovation & research across the continent. Additionally, over the past three years, Digital Umuganda has worked closely with community health workers to explore how artificial intelligence can support their day-to-day responsibilities and capacity development. The company is currently developing an LLM-powered, voice-based tutoring tool for community health workers, designed to strengthen their knowledge and skills, particularly in the area of mental health. The tool aims to provide accessible, contextually relevant, and continuous learning support in the languages community health workers use and understand. Through the AI Hub partnership, Digital Umuganda plans to use the compute infrastructure to expand its African language dataset collection efforts and train larger, more accurate speech recognition, text to speech, Named Entity recognition and translation models.

Learn more about their work

here