A Turkish team of four teachers, connected by family and marriage, developed an artificial-intelligence language agent called MAHİR to analyze students’ learning paths and help teachers monitor their progress. The project won second place in the “Language Artificial Intelligence Agents for Processing and Tracking Government Documents” competition at TEKNOFEST, after competing against hundreds of projects in Kocaeli.
The team consists of Erzurum Provincial Directorate of National Education literature teacher and project coordinator Zelal Ülker Daştan, her husband, religious culture and ethics teacher Lokman Daştan, her sister, science teacher Gonca Ergül, and Ergül’s husband, mathematics teacher Hakan Ergül. Developing the project took approximately five months of work.
How Does MAHİR Work?
According to the team, MAHİR links the “learning evidence” collected by a student during the educational process to learning outcomes, then produces evidence-based educational analyses and improvement suggestions. The system also compiles data drawn from the framework plans and curricula prepared under the “Turkish Century Education Model” to help teachers assess the student’s current stage and identify the next educational step.
The project aims to address the time required to evaluate tests, analyze learning-related data, and monitor students’ progress. The team describes the system as a tool that supports teachers, not a replacement for them, as it is intended to help teachers create a roadmap for interpreting their students’ levels and making subsequent educational decisions.
What Matters About This Development?
The project’s significance lies in the fact that it comes from a team working within the educational process, whose members say the idea began with family discussions about everyday problems in schools. MAHİR also focuses on the Turkish language and data associated with the local education system, rather than providing a general tool detached from the curricula and established assessment mechanisms.
Lokman Daştan said the new education model focuses on the continuous educational process rather than only the final result, increasing the need to analyze documents and measurements related to students. He estimated that implementing the idea could involve processing two billion documents annually; these estimates were conveyed by the agency from the team and were not supported in the article by additional technical or operational details.
Limitations and Open Questions
The source does not explain whether MAHİR is actually being used across schools, what technical infrastructure it relies on, or how student data are protected and the accuracy of its analyses and recommendations is verified. It also does not mention details about the language model, mechanisms for integrating the system with existing education platforms, or independent test results measuring its impact on teacher and student performance. Therefore, winning at TEKNOFEST indicates that the idea stood out in the competition, but it does not yet prove that the system is ready for broad deployment.
The team intends to continue developing the project, while its members believe that their approximately 19 to 20 years of educational experience helped them identify the problem MAHİR is attempting to address.