Teaching and learning of foundational math have emerged as concerns in India, and might be connected with the problems of large classroom sizes and inappropriate curriculum resources. To address these problems of practice, ABC School explored the use of blended learning in three Grade 2 mathematics classrooms (n = 108), while a control group of three Grade 2 classrooms (n = 113) received traditional instruction and curricular resources. A quasi-experimental mixed-methods design was used to determine the differences in teaching, student engagement, and learning after eight months. Standardized test scores showed a significantly higher increase in the performance of the treatment group compared to the control group. Also, the treatment group experienced higher levels of learner-centered pedagogy, measured in terms of the extent of differentiated instruction, timely and precise feedback, and student engagement. This study contributes to the small pool of international experimental studies on blended learning in elementary grade classrooms, and supports a growing belief among researchers that the approach holds promise for school reform in developing countries.