Ultimate access to all questions.
You are working for an online travel agency that provides advertising solutions. Your task is to predict the most suitable web banner for a user to view next, with security as a top priority and a model latency requirement of 300ms@p99. Given thousands of banners in the inventory and exploratory analysis showing navigation context as a strong predictor, what is the simplest way to configure the prediction pipeline? Consider the following constraints: the solution must be cost-effective, comply with data protection regulations, and scale efficiently with increasing user traffic. Choose the best option that meets all these requirements.
Explanation:
The simplest and most suitable solution that meets the latency, security, cost-effectiveness, compliance, and scalability requirements involves using services that ensure fast data retrieval and a managed model deployment. Cloud Bigtable is recommended for its high-throughput, low-latency capabilities, making it ideal for real-time context retrieval. App Engine serves as a scalable gateway, and AI Platform Prediction simplifies model deployment. This combination effectively addresses the need for performance, scalability, and simplicity. Option A is also considered for scenarios where the simplicity of deployment is prioritized, but it may not fully meet all the scalability and data retrieval speed requirements as effectively as option C.