Distributed queuing system, where we can form a cluster with n no of nodes. One master and  n -1 slaves, dockerising total codebase and creating jobs. Master will allocate jobs to slaves, and when any slave went to bad state master will re-schedule that job to another slave. Master will scan and preserve slaves state in database.

Based on user request with company details, we will be mapping with Millions of records with closest match and tie breaking rules. Time complexity is major problem.