Identify Project Resources (e.g., People, Hardware, Contractors) PMI-CPMAI Practice Questions (Page 5)

Questions 21–25

1. 21foundation · easy

   A project manager is creating a timeline for an AI project. The model training phase cannot start until the data preparation phase is complete. What is this type of dependency called?

   • AFinish-to-start dependency
   • BStart-to-start dependency
   • CFinish-to-finish dependency
   • DStart-to-finish dependency
   Reveal Solution Discussion Select an answer first
2. 22application · medium

   A healthcare startup is developing a deep learning model to detect anomalies in medical images. The team will experiment with large image datasets and then deploy the model for real-time inference on edge devices in clinics. The project has a limited budget for cloud GPU usage. Which hardware strategy best balances experimentation needs with deployment constraints?

   • AUse only cloud GPU instances for training and also run inference in the cloud from the clinics.
   • BPurchase a dedicated on-premises GPU server for both training and deployment to avoid cloud costs entirely.
   • CRent a small GPU instance for training and deploy the model on the same instance type in the cloud.
   • DUse cloud GPU instances for training and experimentation, then optimize the model for deployment on edge devices with limited compute.
   Reveal Solution Discussion Select an answer first
4. 23application · medium

   A retail company is planning an AI project to forecast store-level inventory demand. The team will need to clean and integrate data from multiple legacy systems, build and tune forecasting models, and deploy the models to a cloud platform. The company has a small in-house data team with strong statistical skills but no experience with modern ML frameworks or cloud deployment. Which combination of roles and skills should the project manager prioritize when staffing the project?

   • AHire additional data scientists with deep learning expertise and let them handle data cleaning and deployment as part of their role.
   • BContract a full external team to build the entire solution, keeping the in-house statisticians only for validation of the final model.
   • CAssign the in-house statisticians to learn cloud deployment on the job and hire a single data engineer to support all data tasks.
   • DHire a data engineer to handle data pipelines and a cloud ML engineer to manage model deployment, while the in-house statisticians focus on model development.
   Reveal Solution Discussion Select an answer first
5. 24expert · hard

   A bank is launching an AI project to detect fraudulent transactions. The project requires expertise in fraud domain, machine learning, and real-time data processing. The bank has a strong IT team but no data scientists. The project has a tight deadline and a limited budget for external hires. Which approach best addresses the skill gaps while managing cost and timeline?

   • AHire a single senior data scientist with fraud detection experience and train internal IT staff to support model deployment.
   • BOutsource the entire project to an external vendor, including model development and deployment.
   • CHire a team of three data scientists with different specialties to cover all aspects of the project.
   • DTrain the existing IT team in data science through an intensive bootcamp before starting the project.
   Reveal Solution Discussion Select an answer first
6. 25foundation · easy

   A project manager is allocating budget across project phases. Which phase typically requires the most data engineering effort and thus a larger allocation of personnel?

   • AProject closure
   • BModel training
   • CData preparation
   • DDeployment
   Reveal Solution Discussion Select an answer first