Social networking platforms in general recommend new connections to their users. Existing approaches towards these recommendations mostly rely on finding the erstwhile snapshot of a Social Networking Graph (SNG). These recommendations are also made on the basis of immediate profiling of users or their triadic closure, without considering connections that may suit indirectly due to overlapping interests. Therefore, in pursuit of proposing a robust scheme for the same, through this work, we model our approach using a powerful graph clustering algorithm known as SNN-DBSCAN. The connections are suggested on the basis of individual engagement of users in categories namely: \( social, political, education, sports \, \& \, entertainment, health \, \& \, lifestyle\) as against any geometrical intuition. The novelty of this work lies in recommending \(k^{th}\) layer connections for a user by leveraging an intelligent link prediction technique. Contrary to the state-of-the-art policies, our approach introduces this layer-wise priority-based suggestions with greater reliability.