Your company manages an ecommerce website where users can add items to their cart and receive recommendations for additional products they might be interested in. You developed a machine learning model that provides these recommendations in near real time. The current workflow is designed as follows: 1. The website will send a message with relevant user and cart data to Pub/Sub and then awaits a message back from Pub/Sub containing the prediction. 2. The predictions are stored in BigQuery for analytics and reporting purposes. 3. The machine learning model itself is stored in a Cloud Storage bucket, where it will be updated frequently to ensure accuracy and relevance. Considering the goal to minimize prediction latency and reduce the effort required to frequently update the model, how should you reconfigure the architecture?