A business analyst at a retail company wants to build a customer churn prediction model without writing Python code. They have customer transaction data stored in BigQuery. Which GCP service should they use?
Gcp Ml Design Low Code
Question · Easy
A marketing team wants to analyze customer sentiment from product reviews using a pre-trained NLP model without managing infrastructure. Which solution meets this requirement with minimal operational overhead?
Introduction & Exam Overview
0.1 Exam Overview, Blueprint & Certification Logistics
Study guide
A data scientist needs to create a classification model in BigQuery ML to predict whether a customer will purchase based on their browsing history. Which model type should they use?
Tech Exam Prep Podcast
Listen on Spotify
Podcast
A company wants to build a conversational AI agent to handle customer support inquiries without writing code. Which Vertex AI feature should they use?
Question · Medium
An e-commerce company wants to generate product descriptions using generative AI. They need to customize the output tone to match their brand voice. Which approach should they use?
A data analyst needs to create a demand forecasting model for retail inventory planning. The data is already in BigQuery with historical sales and date features. Which is the most efficient approach?
0.2 Study Strategy, Scenario Mechanics & 8-Week Roadmap
A financial services company wants to extract entities like company names and monetary values from loan documents. They need high accuracy and want to avoid training custom models if possible. Which solution is most appropriate?
A data science team needs to share and version control their ML features across multiple projects to ensure consistency. Which Vertex AI service should they implement?
Gcp Ml Collaborate Data Models
A data engineering team needs to preprocess 30 TB of raw clickstream JSON into clean feature tables every night, using custom PySpark code. Which Google Cloud service is the best fit for this batch workload?
A company has trained multiple versions of a fraud detection model using Vertex AI. They need to track which model version is in production and compare performance metrics. What should they use?
Section 1: Architecting Low-Code AI Solutions (~13%)
1.1 BigQuery ML: Model Types, SQL Syntax and In-Database Evaluation
A team is developing an ML pipeline using Vertex AI Pipelines. They need to ensure reproducibility by tracking the exact dataset version used for each training run. Which approach should they implement?
A data science team needs to collaborate on Jupyter notebooks with version control and share compute resources. Which solution is most appropriate?
An ML team wants to implement CI/CD for their model training pipeline. They need to trigger retraining when new data arrives and run tests before deployment. Which tools should they use?
A team needs to serve real-time features to their production model with low latency while ensuring training-serving skew is minimized. What architecture should they implement?
1.2 Feature Engineering and Feature Selection in BigQuery ML
A company needs to share a trained model with an external partner while maintaining control over model access and versioning. Which approach is most secure?
A team is implementing feature engineering pipelines. They need to transform categorical variables into embeddings and share these across multiple models. Where should they store the embedding mappings?
A data scientist needs to choose between AutoML and custom training for a customer segmentation project with limited labeled data. Which approach is more appropriate?
Gcp Ml Build Prototype Production
An ML engineer needs to train a TensorFlow model on a large dataset that does not fit in memory. They want to use distributed training across multiple GPUs. Which Vertex AI feature should they use?
1.3 Generating Predictions with BigQuery ML
A team needs to optimize hyperparameters for a gradient boosting model. They want to search efficiently across a large parameter space. Which approach should they use?
A data scientist is working on an image classification project with only 500 labeled images. They need good accuracy with limited training data. Which technique should they apply?
A team needs to automatically find the optimal neural network architecture for their computer vision task. Which Vertex AI capability should they use?
An ML engineer needs to evaluate a binary classification model beyond accuracy due to class imbalance. Which metrics should they prioritize?
1.4 AutoML on Agent Platform: Tabular, Image, Text and Video Training
A financial services company needs to explain why their loan approval model denied a specific application to comply with regulations. Which technique should they use?
A data science team is building a recommendation system. They have user interaction data and item metadata. Which Vertex AI approach is most appropriate?
A team needs to preprocess image data (resize, normalize, augment) before feeding it to their training pipeline. They want this to scale automatically. Which solution should they use?
An ML engineer is debugging a model that performs well on training data but poorly on validation data. What is the most likely issue?
1.5 Fine-Tuning Gemini Models from BigQuery
A team needs to train a time series forecasting model with multiple related time series (hierarchical forecasting). Which approach should they use?
Question · Hard
A data scientist needs to experiment with different feature engineering approaches and track which combinations yield the best model performance. Which tool should they use?
A team is training a large language model and needs to use techniques to reduce memory usage and enable larger batch sizes. Which technique should they apply?
A company needs to evaluate multiple algorithms (Random Forest, XGBoost, Neural Network) for their fraud detection task. Which Vertex AI approach efficiently compares these options?
1.6 Model Garden: Evaluating and Selecting the Right Model
An ML engineer needs to implement cross-validation for a model trained on time series data. Which approach is most appropriate?
A company needs to deploy a model for real-time predictions with low latency requirements (<100ms). Which deployment option should they choose?
Gcp Ml Scale Deploy
An ML team needs to deploy a new model version but wants to validate it with a small percentage of traffic before full rollout. Which deployment strategy should they use?
A team needs to process millions of records for inference overnight. They want cost-effective batch processing. Which solution is most appropriate?
1.7 Industry-Specific AI APIs: Document AI, Vision, Translate and Speech
An ML engineer needs to reduce model size by 75% for edge deployment while maintaining acceptable accuracy. Which technique should they apply?
A deployed model is experiencing high latency during peak traffic hours. The team wants automatic scaling based on request volume. How should they configure the endpoint?
After deploying a new model version, the team discovers it performs worse than expected. They need to quickly revert to the previous version. What should they do?
A team needs to deploy an LLM for a chatbot application but the full model is too large for their budget. They want to serve it efficiently. Which approach should they consider?
Loading more…