The rapid growth of e-commerce has given consumers an overwhelming number of choices. To help navigate this, our study developed a recommendation system to provide personalized product suggestions. Using the Amazon dataset, we applied collaborative filtering techniques, particularly singular value decomposition (SVD) alongside matrix factorization imputation to discover hidden patterns in customer behaviour. This approach allows the system to predict user preferences and provide personalized product recommendations. Through extensive data pre-processing, hyperparameter tuning, and comprehensive evaluation using metrics such as RMSE, MAE, confusion matrix, precision, and recall, the system achieved an RMSE of 0.7625 and an MAE of 0.5141, indicating its accuracy in predicting user ratings. A user-friendly interface enables seamless interaction and real-time recommendations. The relevance of the achieved results lies in their improvement over baseline methodologies, providing enhanced predictive accuracy and personalized user experiences.