This research studied 1400 anonymous learners from 2007 to 2018 in a biotechnology engineering program in France. The population represents 79% and 21% of females and males, focusing on their first two years over five years to obtain an engineering degree. This cohort has not yet been explored. We employed unsupervised machine learning (ML) methods to identify diverse students’ profiles, while supervised ML methods were used to examine correlations between students’ grades and course content. In addition, we cross-referenced student evaluations of the courses with the course content. Our analysis revealed (1) the relationships between courses, (2) identified three distinct learner subgroups, and (3) extracted student evaluations using sentiment analysis. These findings revealed the predictive factors of student performance in science-related subjects and offered a detailed evaluation of course structures based on students’ feedback. By linking quantitative data with qualitative feedback, this analysis contributes to a better understanding in higher education, providing future curriculum developments and teaching strategies in biotechnology engineering education. However, student success at the start of higher education is complex and multi-factorial. Educational reforms should be considered with more precise parameters to understand their further impact on students’ performance.