- 2025-08-07
- Eduardo Rivera, Yitong Tseo, Christian Beron, Rodrigo Vargas Sainz
Between 2020 and 2025, a seed of innovation planted by MIT’s Global Startup Labs (now part of Global Teaching Labs) in Uruguay became an important milestone, marking one of the first concrete steps toward applying machine learning and artificial intelligence to public policy in Uruguay. What began as an educational experiment has since become a blueprint for impactful, cross-sector collaboration.
Laying the Groundwork
In early 2020, Uruguay became one of the first countries in Latin America to host MIT Global Startup Labs (GSL), a hands-on innovation and entrepreneurship program that officially became part of MISTI’s Global Teaching Labs in 2023. Led by MIT student instructors, the program brought together a multidisciplinary cohort of university students and young professionals in Montevideo for intensive training in machine learning, software development, and business modeling.
Delivered in partnership with the Universidad Tecnológica del Uruguay (UTEC), the program was designed to build a strong foundation in data science while fostering a culture of innovation and social responsibility.
“Our program in Uruguay was designed to empower students to use new AI technologies to address local challenges,” says Eduardo Rivera, managing director of MISTI Uruguay. “We wanted participants not only to understand how to build machine learning models, but also to think critically about how these technologies can serve their communities. At the same time, MIT student-instructors are able to learn and apply their knowledge in new social, economic and cultural contexts.”
Building Capacity and Long-Term Partnerships
Following the success of the 2020 program, MIT’s engagement with Uruguayan institutions deepened. Faculty and staff continued working with UTEC to expand access to technical education, including the MITx MicroMasters in Data Science and Statistics, which was integrated as part of UTEC’s Master’s program in Data Science.
During the pandemic, MISTI pivoted to remote engagement, allowing MIT students and local partners to collaborate on virtual projects, hackathons, and training sessions. These efforts laid the foundation for more advanced applications of AI in the public sector.
From Training to Transformation
A standout example of this collaboration's impact emerged after the 2024 edition of the program. Six MIT instructors worked with around 35 local participants on capstone projects addressing real-world challenges. One such project, led by students Christian Beron Curti and Rodrigo Vargas Sainz, and advised by MIT PhD student Yitong Tseo, explored how machine learning could improve Uruguay Crece Contigo, a national social assistance program run by the Ministry of Social Development (MIDES). “When I first heard about the project idea – it struck me that yes, this is probably the best use of ML I can imagine”, reflects Yitong. “It was a privilege to prototype together during the GSL workshop and continue the work with Christian and Rodrigo even after it concluded; any and all positive impact of the work stands as direct testament to their creative problem-solving and eagerness to make a positive impact.”
The program supports pregnant women and children under the age of four living in extreme vulnerability. Using an anonymized dataset of over 15,000 cases, the team developed predictive models, including gradient-boosted trees, neural networks, LSTMs, and ensemble methods, to identify high-priority cases based on past intervention data.
Their objective was clear: reduce human bias in decision-making, improve prioritization efficiency, and allow technical staff to focus on service delivery rather than manual screening.
The project, detailed in a 2024 preprint on arXiv, used fairness-aware machine learning techniques and synthetic sampling methods to improve both accuracy and equity. As of mid-2025, the model is in the process of being evaluated by MIDES as a decision-support tool. While initial tests have shown promising results, final adoption and full deployment are still under construction.
Reflecting on the experience, Christian Beron Curti notes: “Working with Yitong was an incredibly valuable experience: beyond his technical expertise, what stood out was his kindness and ability to listen. Applying machine learning to a social program demonstrated how timely and valuable it is to incorporate these tools into public policy. While AI is rapidly advancing in many sectors, there is still enormous potential — and need — to innovate in the social domain. This is perhaps one of the most important challenges ahead.”
Rodrigo Vargas Sainz also shared his reflections on the project, emphasizing both the technical challenges and the human dimension of working with social data: “For me, this project was a turning point. It came from a genuine curiosity about how artificial intelligence could be applied to social problems, and it confirmed that these tools, when designed responsibly, can empower programs to better serve those who need it most. Collaborating with MIT instructors and applying advanced techniques to a real policy problem was not only an extraordinary learning experience but also a reminder that technology must always stay connected to human needs.”
A Broader Shift Toward Responsible AI
This work in Uruguay mirrors a broader global trend toward responsible, data-driven policymaking and highlights the vital role of academia in driving public-sector innovation. As AI becomes more embedded in daily life, now used by over two-thirds of the world’s population, the need for local capacity, ethical design, and critical thinking becomes increasingly urgent.
Uruguayan leaders, including software entrepreneur Nicolás Jodal, have underscored the importance of investing in AI talent not just through technical training, but also by nurturing a culture of experimentation and ethical awareness.
A Model for Global Engagement
What began as a classroom initiative has evolved into a case study in long-term, cross-sector collaboration. For MIT, it’s a strong affirmation of the Institute’s global mission.
Between 2020 and 2025, the GTL Uruguay initiative grew from a short-term learning experience into a sustainable talent pipeline, demonstrating how MIT’s global programs can help shape more equitable and effective public policy through partnership, knowledge-sharing, and applied innovation.