SHANGHAI, Oct. 25, 2022 /PRNewswire/ — Oct 13th-14th, the 2022 AWS Summit China with the theme of “Build with Freedom, Explore the Infinite” (official Chinese name: 自由构建、探索无限) kicked off. Tech experts and enterprises gathered at the summit to share their advanced practices and insights in cloud computing from a global perspective, where Shoplazza‘s CTO, Bing Xia, gave a speech about its AI-powered recommendation technology for cross-border eCommerce.

Bing at 2022 AWS Summit China
Bing at 2022 AWS Summit China

The summit hosted by Amazon Web Services (AWS) is its largest annual technology event in China and a benchmark in the global cloud computing field. As a globally leading shopping cart SaaS company, Shoplazza was invited to the summit to share its excellent practices with advanced product features and technology. Shoplazza has provided 360,000+ merchants worldwide with various enterprise-level solutions such as well-crafted eCommerce store themes, automated order management and operations, multi-channel marketing campaigns, and secure and easy-to-follow payment processes.

In recent years, branding has played a vital role in eCommerce. More and more cross-border merchants have engaged in the DTC model and transferred to a refined business operation and management to grow the business. And the “AI-powered recommendation” is one of the crucial features to implement in the online store.

“On a global scale, AI-powered recommendation is an essential technology to integrate for improving sales revenue and customers’ shopping experience.” Bing, CTO of Shoplazza, said, “But it is tough for small to medium-sized businesses (SMBs) to implement the AI solutions designated for large-sized companies. SMB merchants have to put extra effort and time into manually optimizing the product recommendation, which takes a lot of time and reduces operational efficiency. Furthermore, it is not easy for customers to be satisfied with what they see in recommendations. Based on this situation, we aim to provide a technology solution for the AI-powered recommendation that can be applied to all merchants of any size, lower the cost and improve efficiency.”

Bing explained the AI and machine learning system of Shoplazza in 5 stages, including (1) data pull, (2)trigger, (3)pre-ranking, (4) ranking, and (5) re-rank. Shoplazza puts its best effort into data processing to ensure each process runs efficiently and smoothly. However, stand-alone TensorFlow cannot be trained quickly due to the massive amount of data that needs to be processed, such as transaction amount and user access. In addition, distributed machine learning on Spark has several flaws, like slow training and high cost.

To solve the pain points, Shoplazza uses the data flow and computing power of Amazon EMR and Amazon SageMaker and combines all the processes, including sampling, featuring training, and estimation for Shoplazza’s recommendation system data flow. This combined procedure for the data flow enables more efficient data processing and machine learning capabilities. “Based on the samples generated, model training is performed in the SageMaker platform, and the real-time reasoning is deployed. When the user requests recommendations, the prediction can splice user features and product features in real-time and display the feature extraction plug-in to generate samples as requesting SageMaker to get the score and displaying the sorting result to the webpage.” Bing introduced the detailed data flow implementation of AI-powered recommendations with SageMaker.

With Amazon EMR and Amazon SageMaker, Shoplazza eases the management of big data and machine learning infrastructure, integrating the data analysis and machine learning for cost-efficient results.” Up to 300 GB of training data only takes 2 hours to complete, the time is reduced by about 10 hours, and the performance of the product recommendation system is improved by 6 times. To serve our merchants better, we continue to implement and develop AI technology for them to achieve global success.” Bing said.

Bing added, “AI and machine learning are the most transformative technologies in the world today, and they are significant in enhancing customer experience and improving enterprise productivity. In the future, Shoplazza will continue to work with AWS to integrate machine learning from recommendations to product search, content understanding, merchant selection, eCommerce risk control, and other business operations. Technological innovation will solve the merchants’ struggles during store operations.”

As a globally leading eCommerce SaaS platform, Shoplazza will continue to dedicate itself to providing innovative solutions and sharing successful practices for market expansion. At the same time, Shpplazza will continue to stay “Open to More“, perfect its partner ecosystem and collaborate with AWS to empower cross-border merchants to “Start Anywhere” with all-in-one solutions for going global through advanced technology, product innovation and modern data strategy.