Organization: Zindi
Application deadline: June 21st, 2026
ABOUT
The competition, which offers a $5,000 prize pool, invites data scientists, AI researchers, machine learning engineers, and NLP specialists to create models capable of accurately understanding and responding to health-related questions in low-resource African languages such as Luganda, Kiswahili, Akan, and Amharic. The initiative seeks to address the major gap in culturally relevant and accessible healthcare information caused by the dominance of English-language AI systems, particularly in areas of maternal, sexual, and reproductive health where reliable and understandable information is essential for improving health outcomes and informed decision-making.
The Goal of the Challenge
- Understanding health-related questions in supported African languages
- Generating accurate and contextually appropriate responses
- Maintaining fluency in the original language of the query
- Supporting health communication in low-resource linguistic environments
Successful models could eventually support:
- Community health workers
- Rural clinic systems
- Patient education platforms
- AI-powered health assistants
- Digital public health initiatives
Hub for Artificial Intelligence in Maternal, Sexual and Reproductive Health (HASH) includes collaboration between institutions such as:
- Infectious Disease Institute (IDI)
- Makerere University
- Makerere Centre for Artificial Intelligence (MAK-AI)
- Sunbird AI
The consortium focuses on four major research priority areas:
- Maternal health
- HIV
- Sexually transmitted infections
- Adolescent sexual and reproductive health
Evaluation Metrics
The leaderboard score is based on:
- ROUGE-1 F1 — 37%
- ROUGE-L F1 — 37%
- LLM-as-a-Judge — 26%
Understanding the Metrics
ROUGE-1 F1 evaluates:
- Word overlap
- Keyword matching
- Basic content similarity
ROUGE-L F1 measures:
- Sentence structure alignment
- Longest common subsequence similarity
- Fluency consistency
LLM-as-a-Judge evaluates:
- Factual accuracy
- Completeness
- Language appropriateness
- Contextual quality
Prize Pool and Recognition
- First Prize: $2,500 USD
- Second Prize: $1,500 USD
- Third Prize: $1,000 USD
Rules and Participation Guidelines
- Maximum of 5 submissions per day
- Maximum of 50 total submissions
- Open-source tools and languages only
- No automated machine learning tools permitted
- Publicly available pretrained models allowed
- Code sharing outside teams prohibited
Why This Challenge Matters
- Supporting underserved communities
- Improving health literacy
- Expanding multilingual healthcare access
- Strengthening locally relevant AI innovation
How to Apply
Sign In the link below:
Sign In
For more information
Visit ZINDI official website in the link below:
Multilingual Health Question Answering in Low-Resource African Languages Challenge 🩺 – Win $5 000 USD
About the Author
