ARCHIVES

The increasing demand for real-time personalization in e-commerce has highlighted the need for recommendation engines that are both scalable and efficient. Traditional systems often rely on centralized architectures that challenge such as scalability, latency and safeguarding user data privacy. This study introduces a novel design for a real-time personalized recommendation engine that operates within a distributed microservices framework by integrating multiple approaches such as collaborative filtering, content-based filtering, and reinforcement learning while also addressing privacy concerns using federated learning and differential privacy techniques. At the same time, by leveraging real-time data streaming and advanced caching mechanisms, the proposed system delivers low-latency recommendations and makes it highly suitable for large-scale e-commerce applications. Experimental results indicate that this approach significantly outperforms conventional centralized systems in scalability, accuracy of recommendations and response times.