Tool Selection Based on AI Project Requirements AIE Practice Questions (Page 5)

Questions 21–25

1. 21application · medium

   A mid-sized manufacturing firm wants to adopt an AI tool for predictive maintenance. They have a small IT team, a limited budget, and a strong preference for avoiding vendor lock-in. They also need to integrate the tool with their existing on-premises ERP system. Which evaluation criterion should be weighted most heavily when selecting the tool?

   • AIntegration capabilities with the existing ERP system
   • BLargest community support
   • CLowest upfront cost
   • DEase of use for the small IT team
   Reveal Solution Discussion Select an answer first
2. 22foundation · easy

   An AI project requires processing 10 TB of unstructured text data, needs GPU acceleration for model training, and must comply with data residency regulations. Which requirement category is most directly addressed by choosing a GPU-enabled training environment?

   • ABusiness requirement
   • BCompliance requirement
   • CFunctional requirement
   • DTechnical requirement
   Reveal Solution Discussion Select an answer first
4. 23foundation · easy

   A project requires real-time inference on edge devices with limited computational resources. Which tool selection is most appropriate?

   • AA cloud-based inference service
   • BA lightweight inference engine optimized for edge deployment
   • CA large-scale distributed training framework
   • DA data preparation tool
   Reveal Solution Discussion Select an answer first
5. 24expert · medium

   A logistics company is building an AI system to optimize delivery routes. They already use a cloud-based ERP system and a separate on-premises fleet-tracking database. The AI team wants to use a modern ML platform, but the data is split across these two systems. The company needs a unified view of the data for training. Which approach best addresses the interoperability challenge?

   • AUse an integration tool to create a data pipeline that synchronizes the on-premises fleet data with the cloud ERP data into a unified data lake, then use the ML platform on that data lake.
   • BExport the data from both systems to CSV files and manually combine them for each training run.
   • CMigrate the on-premises fleet database to the cloud ERP system to consolidate all data in one place.
   • DHave the ML platform connect directly to both the cloud ERP and the on-premises database at training time.
   Reveal Solution Discussion Select an answer first
6. 25expert · hard

   A financial firm is building a real-time trading model. The model requires ultra-low latency (under 10 ms) and the data is highly sensitive. The firm has a small on-premises data center but also uses a public cloud for other workloads. Which deployment strategy best meets the latency and data privacy requirements?

   • ADeploy the model on-premises with dedicated GPU servers and local data storage.
   • BUse a hybrid approach: train on-premises but deploy in the cloud for scalability.
   • CUse a cloud-based serverless function with a regional endpoint.
   • DDeploy the model in the public cloud with a private connection to the on-premises data center.
   Reveal Solution Discussion Select an answer first