Wastewater generated during electroplating and surface coating metal finishing processes of railcar manufacturing contains pollutants hazardous to the environment leading to depletion of our resources. Previous studies developed advanced methods such as membrane filtration, chemical precipitation, and electrochemical oxidation for the treatment of metal-finishing wastewater, but the water quality of the recycled effluent remains a huge challenge. This study seeks to use an anomaly detection approach that utilizes Variational Autoencoders (VAEs) to detect impurities with high accuracy thereby boosting closed-loop real-time effluent quality in wastewater recycling. Adding to this proposed system, pollutants from closed-loop treated wastewater are also identified and recycled to reduce wastewater discharge and conserve resources. Findings from this study showed that the integration of VAEs helps detect abnormal variations in wastewater quality parameters while allowing early intervention to safeguard system stability. Through this proactive approach, the rail car manufacturing process attains the best standards of recycled water quality besides cutting wastage of critical resources as well as negative effects on the environment making it a sustainable process.