Get Your CertDatabricks Certified Machine Learning ProfessionalPlayCompete
‹ Databricks Certified Machine Learning Professional tests

Databricks Certified Machine Learning Professional practice questions

30 questions in 3 tests. Open one to answer it on its own, or take a whole test.

Test 1

Take Test 1 ›
  1. A retailer needs a separate demand-forecasting model for each of its 2,000 stores. The training data is a 300-million-row Delta table with a store_id column, and any single… · Scaling and Tuning
  2. A team is moving a tuning notebook from Hyperopt SparkTrials to a cluster running Databricks Runtime 17.3 LTS ML. The new setup must meet all of these requirements: - Trials run… · Scaling and Tuning
  3. An engineer builds a Spark ML Pipeline with three stages: an Imputer that fills missing values with column means, a StandardScaler, and a LogisticRegression estimator. The… · Model Development Using Spark ML
  4. A churn model reaches 0.95 AUC offline but only 0.71 in production. The labels DataFrame has customer_id, event_ts, and churned. Features come from the Unity Catalog feature… · Advanced Feature Store Concepts
  5. An organisation uses three Unity Catalog catalogs: dev, staging, and prod. Version 7 of staging.ml.churn has passed all validation. Both the production batch job and the… · Model Lifecycle Management
  6. An ML team manages its weekly retraining job with a bundle (Declarative Automation Bundles, formerly Databricks Asset Bundles). databricks.yml defines a dev target with mode… · Environment Architectures
  7. A credit model's endpoint logs requests to an inference table. A job unpacks the logs into a Delta table, and an inference profile (Lakehouse Monitoring, now called data… · Drift Detection and Lakehouse Monitoring
  8. A team links a Databricks job to the Unity Catalog model prod.ml.fraud as its MLflow deployment job. The job has three tasks: evaluation, approval, and deployment. The job-level… · Automated Retraining
  9. A streaming pipeline writes transaction-velocity features to the Unity Catalog feature table ml.features.card_velocity. The fraud model's serving endpoint looks up these features… · Advanced Feature Store Concepts
  10. A Model Serving endpoint named recs sends 100% of its traffic to version 12 of a recommendation model. The team wants a canary rollout of version 13 with these requirements: -… · Deployment Strategies

Test 2

Take Test 2 ›
  1. A scikit-learn regressor is served behind a real-time endpoint. Before scoring, raw requests need a currency conversion applied to two fields; after scoring, predictions must be… · Advanced MLflow
  2. An engineer wraps a retrieval model in a custom mlflow.pyfunc.PythonModel. The class holds a handle to an external service client created in __init__. Calling… · Advanced MLflow
  3. A Spark ML CrossValidator tunes a classifier over a training DataFrame built from four joins and several filters across Delta tables. The Spark UI shows the same joins and file… · Model Development Using Spark ML
  4. A fraud model needs a feature defined as the ratio of the current transaction's amount to the cardholder's 30-day average, plus the minutes elapsed between the request timestamp… · Advanced Feature Store Concepts
  5. A retraining pipeline is deployed with a bundle and its code lives in a Git repository. The team wants two things: a pull request must fail within seconds if someone breaks a… · Testing and CI/CD
  6. A data profile runs on an inference table with a baseline table configured. The team needs two things beyond the built-in metrics: a revenue-weighted error figure computed from… · Drift Detection and Lakehouse Monitoring
  7. A custom model is served on a Databricks endpoint. For later drift analysis the team needs every request payload and the corresponding response landed in a Unity Catalog Delta… · Drift Detection and Lakehouse Monitoring
  8. An organisation is designing promotion for a retraining pipeline across dev, staging and prod workspaces. Production data is governed and readable only from the production… · Environment Architectures
  9. A nightly retraining job must stop automatically, before any registration step, whenever the newly trained classifier scores below an F1 of 0.80 on the holdout set. No person is… · Automated Retraining
  10. An internal endpoint serves a few requests an hour during business hours and none overnight, and it was created with scale to zero enabled. Users report that the first request… · Deployment Strategies

Test 3

Take Test 3 ›
  1. A training dataset is 8 GB and the cluster's worker nodes have 128 GB of memory each. The team is comfortable with scikit-learn and is deciding whether to rewrite their training… · Model Development Using Spark ML
  2. A Spark ML LogisticRegression is trained on a 2-billion-row table in which 0.5% of rows are positive. The fitted model predicts the negative class for almost every record. The… · Model Development Using Spark ML
  3. A PyTorch image model takes about 30 hours to train on one GPU. The cluster has four GPU workers. The team wants to train a single model across all four GPUs, keep their existing… · Scaling and Tuning
  4. An auditor asks which exact table and table version produced version 9 of a production model. The training runs log hyperparameters, metrics and the model artifact, but nothing… · Advanced MLflow
  5. A retraining job must read from and write to dev_catalog in development and prod_catalog in production. Everything else about the job (tasks, cluster spec, schedule) is… · Environment Architectures
  6. A serving endpoint is created under a service principal to serve prod.ml.fraud. The service principal already holds USE CATALOG on prod and USE SCHEMA on prod.ml, but the… · Model Lifecycle Management
  7. Labelled outcomes are appended to a Unity Catalog table at irregular intervals, sometimes twice a day, sometimes not for three days. A retraining job currently runs on an hourly… · Automated Retraining
  8. A declarative pipeline computes features from an upstream feed that occasionally emits a negative amount or a null customer_id. These rows must not reach the feature table, the… · Testing and CI/CD
  9. A production retraining job runs a notebook stored in an engineer's personal workspace folder. Last week someone edited the notebook while a run was in progress, and the team… · Environment Architectures
  10. A transformer-based model that was trained on GPUs is deployed to a Model Serving endpoint created with the default configuration. Latency per request is an order of magnitude… · Deployment Strategies