Earthquake is one of the most devastating natural hazards which can lead to catastrophic losses both in terms of life and economy. Seismic risk assessment provides the platform for improving disaster management policies and to prioritize the limited resources of the country. The primary component of seismic risk is to identify seismic vulnerability of existing buildings and other infrastructure. Collection of Building Inventory Data (BID) requires extensive field surveys which is cumbersome and time consuming. This study captures growth and development pattern of the city for clustering utilizing GIS and ML with reduced time and efforts. Essential parameters viz. built-up density, building age, demography, street intersection density, tree density and important structures have been identified from various sources such as GIS tools, census data, satellite imagery and plugins. The above-mentioned parameters were combined using Analytical Hierarchy Process (AHP) and the results were verified using predictive machine learning models. For each cluster, building level characterization and vulnerability assessment has been carried out using HAZUS methodology. Finally, seismic risk maps for Nagpur city have been developed in 100 m × 100 m grids which suggests 1.43% population will be affected (injury and casualty) for MCE level hazard. Moreover, the proposed GIS and ML based methodology helps to identify clusters with acceptable accuracy and minimal human error, which further simplifies the seismic risk assessment procedure as well as significantly reduce the computation time.