A recommendation system, often called a recommender system, is a kind of artificial intelligence (AI) algorithm that suggests or recommends products to both current and potential clients based on big data. It makes use of data to anticipate, target, and pinpoint what customers are looking for from an ever-expanding range of options. To identify them, a variety of indicators may be employed, such as past purchases, search history, demographic data, and other factors. It helps the users locate products and services for people that they are unable to locate to help them. At the initial level with a new customer, due to lack of knowledge, the Recommender System (RS) has a cold-start issue while making suggestions. Upon registering with the system, new users do not have access to any history of their choices or interactions. Without this information, the system is unable to offer customized recommendations. Furthermore, a newly introduced object to the system has no pre-existing interactions or preferences with other things. This poses a challenge for the algorithm to make recommendations based on user preferences. The cold start issue can limit the effectiveness of recommendation systems as they struggle to provide suggestions due to lack of information about individuals and products. This may result in users not returning to the system leading to a user experience. Recommendation systems adopt strategies like content-based filtering, collaborative filtering, hybrid systems, knowledge-based systems and demographic data to overcome the cold start problem. In this paper a method combining Cosine Similarity (CS) and Matrix Factorization (MF) is proposed as a solution, for addressing the cold start problem and further resolving sparsity challenges.