Undergraduates studying Computer Science (CS) and engineering need to hone their problem-solving abilities, which are especially crucial when tackling programming challenges involving basic strategies. These abilities should ideally be cultivated earlier in the programming learning process, especially in the preliminary few days or weeks of a Novice Student’s (NS) first course. The growth of programming skills for total novices depends on their ability to solve problems and think logically. The primary element that may impact NS computing abilities is cognitive impairment and lack of cognition impairs students’ ability to solve arithmetic and problem-solving issues. For many NS, Introductory Programming (IP) might be complicated. Additionally, these courses have a high dropout and failure rate. Since it enables human-AI relationship towards analytical guidance, where monitors will be instructed on how to assist and help NS where early interference is vital for predicting NS performance at an early stage is one possible solution to this issue. In the current study, a Prediction Model (PM) has been created especially to predict how well first-year CS will perform in the first challenging course in their field. In this regard, the study used data gathered during the first two weeks of IP courses delivered to a total of 2372 students to apply suggested Bayesian Optimization Hyperparameter Tuning (BOHT) for early performance prediction. The main aim of this research is to progress a way to identify “at risk” students early on and offer advice on how to help them become better programmers. The experimental results showed that the proposed BOHT model produced the best results when compared to other existing models, with an accuracy of 98.3%.