1 Adaptive teaching does not usually rhyme with large number of students in class. Given the small amount of face to face and the disproportionate ratio, teachers have to use their intuition and experience to fine tune teaching strategies and contents. We report in this paper on our experience of using an intelligent teaching assistant system in a large Computer Science undergraduate class. Whilst helping students practice with tailored and immediate feedback, it also captures data so that the teaching staff can be informed in real time of the students' problems and progress. We show how this information is used to adapt the teaching to the current class.