Objective 4.1 Deploy a Model AIP-210 Practice Questions (Page 1)

Questions 1–5

1. 1foundation · easy

   A financial services company needs to generate credit risk scores for millions of customer records each night. The predictions do not need to be available immediately. Which deployment strategy best fits this use case?

   • AReal-time deployment
   • BEdge deployment
   • CBatch deployment
   • DServerless deployment
   Reveal Solution Discussion Select an answer first
2. 2application · medium

   A retail company has a demand forecasting model that runs nightly to predict inventory needs for the next day. The predictions are consumed by a batch process that updates warehouse stock levels. The model's inference time is about 2 minutes per store, and there are 500 stores. Which deployment strategy best fits this use case?

   • AReal-time API deployment with a load balancer to handle concurrent requests
   • BServerless function triggered by a timer that processes one store per invocation
   • CBatch deployment that processes all stores in a scheduled job
   • DEdge deployment on each store's local server
   Reveal Solution Discussion Select an answer first
4. 3application · medium

   An e-commerce company expects a large spike in traffic during a holiday sale. Their ML model for product recommendations is currently deployed on a single VM. What is the best way to handle the expected increase in prediction requests?

   • AUse a load balancer in front of multiple VM instances and enable autoscaling
   • BLimit the number of requests per user to reduce load
   • CDeploy the model on a larger VM with more CPU and memory
   • DMove the model to a batch processing pipeline
   Reveal Solution Discussion Select an answer first
5. 4application · medium

   A team has trained a deep learning model in PyTorch and wants to deploy it to a mobile app for offline image classification. The model is 500 MB and the app needs to run on low-end devices. What is the best approach?

   • AConvert the model to ONNX and then to a mobile-optimized format like Core ML or TFLite
   • BCompress the model using zip and deploy it
   • CDeploy the PyTorch model as-is to the mobile app
   • DHost the model on a cloud API and require internet connectivity
   Reveal Solution Discussion Select an answer first
6. 5foundation · easy

   What is a key metric to monitor for a deployed model's system health?

   • ASize of the training dataset
   • BAccuracy on the training set
   • CCPU utilization
   • DNumber of training epochs
   Reveal Solution Discussion Select an answer first