We’re excited to announce the winners of the UNIDO AfricaRice Quality App Challenge, where participants were tasked with going beyond model development to build fully functioning mobile apps capable of analysing rice quality from images.

This challenge marked an important milestone for Zindi as our first-ever app building challenge, and the quality of the submissions was outstanding. Zindians demonstrated not only strong machine learning skills but also thoughtful app design, offline functionality, and user-friendly workflows, resulting in practical tools that could be used in real-world settings.

🧑‍⚖️ Final Evaluation

While the public leaderboard reflected alignment of demo videos to requirements, all Top 15 submitted apps were human-verified as part of the final evaluation. The judging process reviewed each solution against the challenge requirements, including usability, offline functionality, speed, and clarity of results.

⚙️ Lessons for Future Challenges

One key takeaway is that scoring can be improved, so that public leaderboard scores better reflect final judging and private scores. We’re already working on improvements to ensure stronger alignment in future challenges.

🎉 A Big Thank You

Congratulations to the winners and to all participants who helped make Zindi’s first app building challenge such a success. We’re proud to see the community pushing AI solutions beyond models and into real-world applications. Your challenge certificates are ready to share and download.