We organized hackathons in Ethiopia, Cameroon, and Senegal, bringing together clinicians, researchers, and software developers to co-create AI solutions addressing local health challenges. These hackathons produced innovative tools that we are now refining and integrating into institutional workflows.
Held in August 2025, the CODE2CARE Hackathon brought together over 140 participants from across Cameroon to co-create AI-driven solutions for challenges identified by the Douala General Hospital.
The CODE2CARE hackathon combined a capacity-building phase, featuring hands-on training in Machine Learning, Generative AI, and digital health, with a competition phase, where multidisciplinary teams developed innovative solutions to address real healthcare challenges.
The hackathon resulted in three innovative solutions:
- Patient Feedback and Reminder Management System
- AI-Powered Education and Diagnostic Support (LLMs)
- AI-Powered Blood Bank Stock Monitoring and Forecasting
The Blood Bank solution has since evolved into the LifeFlow & CrimsonFlow platform, which is currently being piloted at the Douala General Hospital through dedicated donor and hospital interfaces for digital blood bank management. An external deployment involving key stakeholders is planned to extend the platform to other health facilities and promote wider adoption across the healthcare system.
Held from April 24 to May 22, 2025, brought together 30 young and motivated AI and data science professionals to tackle two critical data challenges: retrospective record linkage and data anomaly detection. Organized through the collaborative efforts of AHRI and its DSWB Data Science Without Borders (DSWB), Armauer Hansen Research Institute (AHRI) consortium partners, the hackathon aimed to develop robust, scalable, and context-sensitive solutions for fragmented health data systems commonly found in low- and middle-income countries.
Over three intensive weeks two virtual and one-in-person hackathon engaged participants through a structured program featuring a kickoff session, ongoing mentorship from local and international experts, regular check-ins, and final pitching. A total of 15 teams, formed based on complementary skills, competed by designing and building AI-driven models using real-world-inspired datasets. The event emphasized practical innovation, technical rigor, and interdisciplinary collaboration.
The final judging showcased high-quality, deployable solutions that addressed challenges like missing data, quasi-identifiers, and noisy datasets. Winning teams were recognized for excellence in innovation, clean coding, and user interface design, with prizes totaling 200,000 ETB and offers of continued mentorship and research collaboration. Participant and technical team feedback highlighted the event’s strong learning impact, collaborative spirit, and areas for improvement.
