Technical and tactical factors of an educational platform aiming to deal with a large number of students and inequalities when teaching programming for STEM students

Considering the characteristics found in the post-pandemic scenario of public higher education in Brazil, we must address issues related to equity in access to educational resources. In STEM (Science, Technology, Engineering and Mathematics) courses, these differences become striking regarding the conditions and resources for extra-class study that the student can count on. Machine Teaching is a web-based system used in introductory programming classes to support students and instructors and has been used since 2018 in Universidade Federal do Rio de Janeiro. In this work, we present the challenges faced to adapt Machine Teaching to the post-pandemic scenario. Issues of equity in access were addressed in three ways: changes to the system architecture, optimization of page loading times, and actions to improve user satisfaction. Our novel implementation allows students to undertake exercises from any device with internet connectivity, enabling access from low computational power machines. Also, our optimization results decreased up to 80% the web system loading time, allowing slower connections to still access the system. This work contributes to the discussion on the risks and benefits of using digital technologies in education, placing our specific scenario in the big picture of inequality of conditions in connectivity, AI and data analysis in education.