Our publications
- Large language models for frontline healthcare support in low-resource settings — Nature Health, 2026. LLMs remained superior to clinicians and were over 500 times cheaper per response in Rwanda’s health system.
- Waxal: A large scale multilingual African language dataset — arXiv, 2026. A speech dataset covering 21 African languages with ASR and text-to-speech components.
- Human Evaluators vs. LLM-as-a-Judge: Toward Scalable, Real-Time Evaluation of GenAI in Global Health — medRxiv, 2025.
- Global PIQA: Evaluating Physical Commonsense Reasoning Across 100+ Languages and Cultures — arXiv, 2025. A benchmark spanning 116 language varieties.
- Mbaza RBC: Deploying and evaluating an LLM-powered chatbot for community health workers in Rwanda — AAAI 2025 Workshop. 66% preferred text, 97.5% satisfaction with voice.
- Kinyarwanda TTS: Using a multi-speaker dataset to build a Kinyarwanda TTS model — AfriNLP, 2023.
- Multilingual automatic speech recognition for Kinyarwanda, Swahili and Luganda — AfriNLP, 2023. 21.91 word error rate across three East African languages.
- Developing an IoT-based conversational AI recommender assistant for vital sign predicted anomalies — ICMLT, 2023.