Tool Selection Based on AI Project Requirements AIE Practice Questions (Page 7)

Questions 31–35

1. 31application · medium

   A manufacturing company is deploying a predictive maintenance model. The model has been trained and validated. The operations team needs to monitor the model's performance in production and retrain it when accuracy drops. Which tool category is most relevant for this ongoing need?

   • AModel training framework.
   • BModel monitoring and management platform.
   • CData preparation tool.
   • DData visualization tool.
   Reveal Solution Discussion Select an answer first
2. 32foundation · easy

   In which phase of the AI project lifecycle would a team most likely use a tool like Apache Kafka or a data ingestion service?

   • AModel training
   • BModel monitoring
   • CData collection
   • DModel deployment
   Reveal Solution Discussion Select an answer first
4. 33application · medium

   A retail company is starting an AI project to forecast inventory demand. The team has historical sales data in a SQL data warehouse and wants to build and train a model, then deploy it as a batch job that runs nightly. The team is small and has limited ML expertise. Which set of tools best fits the project lifecycle from data preparation to deployment?

   • AUse a data preparation tool to clean the data, a cloud AutoML service to train the model, and a scheduled job to run predictions.
   • BUse a deep learning framework to build a custom model from scratch, then deploy on a Kubernetes cluster.
   • CUse a commercial BI tool to visualize data and a separate open-source library for model training, with no deployment plan.
   • DUse a spreadsheet application for data cleaning, a custom Python script for training, and a manual process to run predictions.
   Reveal Solution Discussion Select an answer first
5. 34expert · medium

   A manufacturing company uses a legacy data warehouse and a custom data pipeline. They want to introduce an AI-based quality-inspection system. The data science team prefers to use a modern ML framework, but the IT department requires that any new tool must integrate with the existing data pipeline without major rework. Which approach best satisfies both the data science team and IT?

   • AUse a different ML framework that is already compatible with the legacy warehouse, even if the data science team is less familiar with it.
   • BReplace the legacy data warehouse with a modern data lake that is natively compatible with the ML framework.
   • CHave the data science team manually export data from the warehouse and load it into the ML framework each time.
   • DSelect an ML framework that can read data directly from the legacy warehouse through a connector, and use a staging area to transform data into the framework's expected format.
   Reveal Solution Discussion Select an answer first
6. 35application · medium

   A government agency is planning an AI project to automate document review. The agency has a strict procurement process that requires any software to be commercially supported and have a clear vendor for accountability. The project also requires the ability to customize the model for specific document types. Which tool selection approach best fits these constraints?

   • AChoose an open-source tool with a large community, and rely on community forums for support.
   • BUse a commercial platform but avoid any customization features to keep the project simple.
   • CSelect a free tool that is widely used in other government agencies, even if it lacks formal vendor support.
   • DSelect a commercial AI platform that offers vendor support and also provides an SDK for custom model training.
   Reveal Solution Discussion Select an answer first