China and Latin America join hands
to advance poverty alleviation;
empower with wisdom
and achieve far-reaching impact through action.
A platform for
cross-cultural exchange, collaboration, and action.
In this spotlight, we’re introducing one of our participating teams
and discover how they are turning
innovative ideas into real-world impact.
On August 14 local time, “2026 Poverty Alleviation Challenge” Brazil Final was successfully held at Universidade Federal do Rio de Janeiro. Focusing on rural education, the Farol team won the silver award with its offline AI education project.
Project Overview
Many rural schools in Brazil suffer from poor internet connectivity and unreliable electricity. Teachers often lack adequate subject-specific training, while students have little access to after-school tutoring at home. Farol (Portuguese for “Lighthouse”) is an offline AI education system built upon Tsinghua University’s open-source OpenMAIC Multi-Agent Classroom. Designed specifically for the realities of rural Brazil, the system supports teachers with lesson preparation and classroom instruction while enabling students to continue learning after school. Offline learning hubs are deployed in easily accessible community locations such as churches and community libraries. Students simply connect to the local Wi-Fi using their own smartphones—without requiring internet access or mobile data—to access educational content. The Portuguese-language cloud platform has already been successfully launched, and pilot schools in the Brazilian state of Minas Gerais (MG) are currently using the system and providing user feedback.
Team Members
ZHANG Henan: School of Software, Tsinghua University; Team Leader; Product Design and Software Development
HAN Lu: Department of Industrial Engineering, Tsinghua University; Hardware Design; Project Documentation
CHIU Yuen Ting: School of Economics and Management (GMBA), Tsinghua University; Business Model and Financing Strategy
Leonardo Vaz: Federal University of Rio de Janeiro (UFRJ); Brazilian Law, Software Development, and Localization
Poliana Cassia: Federal University of Minas Gerais (UFMG); Former K–12 Teacher; Field Research and Community Engagement
Reflections on Winning the Award
There are already plenty of AI education products on the market, but they all share one basic assumption: there has to be internet access. What we want to do is bring these technologies to places where there is no internet at all, so that children and teachers there can benefit from them as well. Interestingly, the offline constraint itself forced us to come up with a lot of creative ideas. Our Brazilian teammate told us that more than 90% of Brazilians belong to a religious community, and that churches are where children actually go after school. The teammate also pointed out that 81% of children already own a mobile phone; what they lack is electricity, internet connectivity, and educational content—not devices. These local realities directly reshaped our product design: the box only needs to provide Wi-Fi and learning materials. Once it is placed in a church, students can simply bring their own phones and start learning.
Looking back, every key decision in this project came from Chinese and Brazilian team members combining the knowledge and perspectives each of us brought to the table. Throughout the competition, the guidance of Professors LI Hongru, SHAO Lei, and PENG Bo was invaluable. From refining the poverty alleviation logic and localization strategy to improving our external narrative, many of our most important ideas took shape through their mentorship. We would also like to thank all the faculty and staff at the Tsinghua University Latin America Center. They organized every aspect of the competition—from the event process and mentor matching to connecting us with resources for local implementation—with remarkable clarity and efficiency, helping us move the project from an initial idea to something genuinely ready for deployment.
Looking Ahead to the Final Round
Our primary goal in the final round is to place the platform directly in the hands of Brazilian teachers and observe how they actually use it—what works well, where they encounter difficulties, and how the system can better support their teaching. We also hope this opportunity will help us identify long-term local partners who are willing to work with us to further develop and implement the project.
