PMLE Architecting Low-Code ML Solutions Practice Test 2 — 15 Questions

Question 1

A healthcare company needs to extract structured data from scanned patient intake forms, including handwritten fields like patient name and date of birth. They require high accuracy and want to minimize custom code. Which Google Cloud solution should they use?

Choose an answer to begin — your selection is scored in the full session.

Accepted Answer

Document AI Custom Extractor. Document AI Custom Extractor is designed for extracting specific fields from documents, including handwritten text, by training on labeled samples. It provides a low-code interface for labeling and training, and can achieve high accuracy for custom entities like patient name and date of birth. Other options either lack structure extraction or are not optimized for handwriting, making Custom Extractor the correct choice.

Answer

Vision API OCR

Answer

Document AI Form Parser

Answer

AutoML Vision

Question 2

A marketing team wants to build a model to predict customer lifetime value (CLV) using tabular data in BigQuery. They have limited ML expertise and want to use a low-code approach. They also need to explain which features contribute most to predictions for business stakeholders. Which TWO approaches should they use? (Choose two.)

Accepted Answer

Use BigQuery ML to train a linear regression model and use the ML.EXPLAIN_PREDICT function to get feature attributions.. The team needs a low-code approach that also provides explanations. BigQuery ML linear regression with ML.EXPLAIN_PREDICT offers SQL-based training and per-prediction feature attributions. Vertex AI AutoML Tables provides automatic model building and global feature importance scores. Both avoid heavy coding and deliver interpretability, making them the best fit for predicting CLV while satisfying stakeholder needs.

Answer

Use BigQuery ML to train a deep neural network (DNN) model and use the ML.PREDICT function to get predictions.

Answer

Use Dataflow to preprocess data and train a custom TensorFlow model with SHAP for explanations.

Answer

Use BigQuery ML to train a boosted tree model and use the ML.FEATURE_IMPORTANCE function to get global feature importance.

Question 3

A retail company wants to forecast daily sales for the next 90 days using historical sales data. They have no ML expertise and want a fully managed service that automatically handles time series modeling, including seasonality and holidays. Which Google Cloud service should they use?

Accepted Answer

Vertex AI Forecast. Vertex AI Forecast is a no-code, fully managed service for time series forecasting. It automatically detects seasonality, holidays, and trends, and allows users to generate forecasts without writing code. The retail company can upload historical sales data and get predictions for the next 90 days. Other options require coding or manual feature engineering, which the company wants to avoid.

Answer

AI Platform Prediction

Answer

BigQuery ML with ARIMA_PLUS

Answer

AutoML Tables

Question 4

A retail company wants to build a model to predict customer lifetime value (CLV) using tabular data with 50,000 rows and 20 features. They have limited ML expertise and want to minimize coding. They also need to be able to explain which features most influence the predictions. Which Google Cloud service should they use?

Accepted Answer

Vertex AI AutoML Tables (now Vertex AI Tabular).. Vertex AI AutoML Tables is designed for tabular data and automates the ML pipeline, from feature engineering to model training and hyperparameter tuning, with minimal coding. It provides feature importance and explainable AI, which is crucial for understanding CLV drivers. Other options either require more coding (BigQuery ML, custom TensorFlow), are not for training (AI Platform Prediction), or are overkill (Cloud TPU). AutoML Tables strikes the right balance of low-code and explainability for this use case.

Answer

BigQuery ML with a linear regression model.

Answer

Cloud TPU with a custom TensorFlow model.

Answer

AI Platform Prediction with a pre-built scikit-learn model.

Question 5

A healthcare startup wants to build a model to predict patient readmission risk using structured electronic health records (EHR) data. They have a small dataset of 5,000 records and need to ensure the model is interpretable for regulatory compliance. They prefer a low-code solution but require feature importance explanations. Which approach should they take?

Accepted Answer

Use BigQuery ML with a logistic regression model and use the ML.EXPLAIN_PREDICT function. BigQuery ML with logistic regression and ML.EXPLAIN_PREDICT offers a low-code, interpretable solution. Logistic regression is suitable for small datasets and provides coefficients that are easy to explain. ML.EXPLAIN_PREDICT returns per-instance feature attributions, which can satisfy regulatory requirements. Other options either require more coding, are not low-code, or do not provide per-instance explanations natively.

Answer

Use AutoML Tables with export to BigQuery and then apply LIME

Answer

Use Vertex AI AutoML Tables and enable feature importance

Answer

Use Vertex AI Training with a custom TensorFlow model and integrate SHAP

Question 6

A marketing team needs to build a model that predicts whether a customer will respond to a promotion. They have a BigQuery table with 5 million rows and 30 features, including numeric and categorical columns. They want to use SQL for training and prediction, and require automatic feature preprocessing such as one-hot encoding and normalization. Which BigQuery ML model type should they use?

Accepted Answer

AutoML Tables with the CREATE MODEL statement using the AUTO_ML option. AutoML Tables is the correct choice because it automatically handles feature preprocessing like one-hot encoding and normalization, which the team requires. It supports binary classification and integrates with BigQuery ML via SQL, allowing training and prediction without manual feature engineering. Other BigQuery ML models either require manual preprocessing or are unsuitable for the task, making AutoML Tables the best fit for the scenario.

Answer

Boosted tree classifier with the BIGQUERY_ML.BOOSTED_TREE_CLASSIFIER function

Answer

k-means clustering with the BIGQUERY_ML.KMEANS function

Answer

Logistic regression with the BIGQUERY_ML.LOGISTIC_REG function

Question 7

A financial services company wants to detect fraudulent transactions in real-time. They have a labeled dataset in BigQuery and want to build a model with minimal coding. They also need to deploy the model for online predictions with low latency. Which approach should they use?

Accepted Answer

Train a BigQuery ML logistic regression model, export it to Vertex AI, and deploy it to an endpoint for online predictions.. Training a logistic regression model in BigQuery ML is low-code and leverages existing data. Exporting the model to Vertex AI and deploying it to an endpoint provides the necessary low-latency online predictions for real-time fraud detection. This pattern avoids data movement and custom coding, aligning with the team's constraints.

Answer

Use BigQuery ML to train a model and use the ML.PREDICT function directly for real-time scoring.

Answer

Use Vertex AI AutoML Tables to train a model and deploy it to an endpoint for online predictions.

Answer

Use Dataflow to stream transactions and apply a custom TensorFlow model for real-time scoring.

Question 8

A media company wants to automatically generate tags for their video content to improve searchability. They have a large archive of videos and want to use Google Cloud's pre-trained AI services to extract labels for objects, actions, and scenes. They also need to extract speech transcripts to generate tags from spoken content. Which TWO Google Cloud services should they use? (Choose two.)

Accepted Answer

Speech-to-Text API. To automatically tag video content, the company needs to analyze both visual and audio components. Video Intelligence API provides pre-trained models for detecting objects, actions, and scenes, and can also transcribe speech. Speech-to-Text API offers robust audio transcription, which can be used to generate tags from spoken words. Together, these services cover the required capabilities without custom model training. Other services either do not process video or are not designed for this purpose.

Answer

Cloud Vision API

Answer

Dialogflow

Answer

Natural Language API

Question 9

A company wants to build a low-code ML solution to classify customer support tickets into categories (e.g., billing, technical, account). They have a dataset of 10,000 labeled tickets and want to minimize coding. They also need to explain why the model made a particular prediction for compliance. Which TWO Google Cloud services or features should they use? (Choose two.)

Accepted Answer

Vertex AI AutoML Natural Language. Vertex AI AutoML Natural Language provides a no-code way to train a custom text classification model on labeled tickets. Vertex AI Explainable AI then provides per-instance explanations for predictions, satisfying compliance. Together, they offer a low-code solution with interpretability. Other options either require coding, do not support custom training, or do not provide explanations.

Answer

BigQuery ML with a logistic regression model

Answer

AI Platform Prediction

Answer

Cloud Natural Language API

Question 10

A healthcare analytics team wants to build a model to predict patient readmission risk using structured data in BigQuery. They have limited ML expertise and need to automatically handle missing values and categorical features. Which BigQuery ML model type should they use?

Accepted Answer

BOOSTED_TREE_CLASSIFIER. BOOSTED_TREE_CLASSIFIER in BigQuery ML is designed for tabular data and automatically handles missing values and categorical features, reducing the need for manual preprocessing. This aligns with the team's limited ML expertise and the need for a low-code solution. It also provides strong predictive performance for binary classification tasks like readmission risk.

Answer

LOGISTIC_REG

Answer

KMEANS

Answer

DNN_CLASSIFIER

Question 11

A healthcare company wants to build a low-code ML solution to predict patient readmission risk using structured electronic health records (EHR) data. They have a large dataset with mixed numeric and categorical features, and they want to minimize coding effort. They also need to explain model predictions to clinicians. Which TWO Google Cloud services should they use? (Choose two.)

Accepted Answer

Vertex AI Explainable AI. The correct combination is BigQuery ML with AutoML Tables and Vertex AI Explainable AI. AutoML Tables enables low-code model training on structured EHR data, while Vertex AI Explainable AI provides the necessary transparency by showing feature importance for predictions. Together, they satisfy the requirements for minimal coding and explainability. Other options either require coding, serve different purposes, or do not directly contribute to building and explaining the model.

Answer

Vertex AI Model Monitoring

Answer

Vertex AI Workbench with custom TensorFlow code

Answer

Vertex AI Feature Store

Question 12

A media company wants to automatically transcribe and analyze customer support calls to identify common issues. They need to extract topics and sentiment from the transcripts. They prefer a low-code solution that requires minimal ML expertise. Which Google Cloud service should they use?

Accepted Answer

Speech-to-Text API followed by Natural Language API. The combination of Speech-to-Text API and Natural Language API is the correct answer because both are fully managed, pre-trained services that require no ML expertise. Speech-to-Text transcribes the audio, and Natural Language API extracts sentiment and entities/topics. This provides a low-code solution that directly meets the requirements. Other options either require custom model training, are not designed for post-call analysis, or are more complex than necessary.

Answer

Vertex AI AutoML Natural Language

Answer

Dialogflow CX with custom intents

Answer

Contact Center AI Insights

Question 13

A financial institution wants to detect fraudulent transactions in real time. They have a labeled dataset of historical transactions and want to build a model using BigQuery ML. They need to handle class imbalance and require model explainability. Which two options should they use to achieve their goals? (Choose two.)

Accepted Answer

Use logistic regression with class weights to handle imbalance.. Using logistic regression with class weights addresses the class imbalance by giving more importance to fraudulent transactions, improving detection. ML.EXPLAIN_PREDICT provides feature attributions for individual predictions, satisfying the explainability requirement. Together, they form a robust solution for real-time fraud detection with BigQuery ML, leveraging low-code SQL capabilities.

Answer

Use matrix factorization for collaborative filtering.

Answer

Use k-means clustering to identify anomalies.

Answer

Use a deep neural network with dropout and early stopping.

Question 14

A retail company wants to predict customer churn using historical purchase data stored in BigQuery. The data includes customer demographics, transaction history, and support interactions. The team is comfortable writing SQL and wants to avoid moving data to a separate environment. Which approach should they take?

Accepted Answer

Use BigQuery ML to create a logistic regression model (LOGISTIC_REG) on the data directly in BigQuery.. BigQuery ML allows the team to build and train a logistic regression model directly on data stored in BigQuery using SQL syntax, without moving data to a separate environment. The LOGISTIC_REG model type is specifically designed for binary classification tasks like churn prediction, and it runs entirely within BigQuery's serverless infrastructure, satisfying the team's requirement to avoid data movement.

Answer

Use the Cloud Natural Language API to analyze customer support interactions and combine results with purchase data in BigQuery.

Answer

Export the data to a CSV file and use Vertex AI AutoML Tables to train a classification model.

Answer

Create a Dataflow pipeline to stream data to Cloud SQL and use Cloud SQL's built-in ML functions.

Question 15

A data scientist needs to train a time-series forecasting model on historical sales data stored in BigQuery to predict future demand. The data has strong seasonal patterns. Which BigQuery ML model type should they use?

Accepted Answer

ARIMA_PLUS. ARIMA_PLUS is the correct choice because it is specifically designed for time-series forecasting in BigQuery ML, handling seasonal patterns, trend decomposition, and automatic hyperparameter tuning. It models autoregressive (AR) and moving average (MA) components with seasonal differencing, making it ideal for historical sales data with strong seasonal cycles.

Answer

MATRIX_FACTORIZATION

Answer

BOOSTED_TREE_REGRESSOR

Answer

K_MEANS