ML lifecycle stages in MLOps - Practice Problems & Coding Challenges

Challenge - 5 Problems

🧠 Conceptual

intermediate

Identify the correct order of ML lifecycle stages

Which of the following lists the ML lifecycle stages in the correct order?

A Model Training, Data Collection, Model Deployment, Model Monitoring

B Data Collection, Model Deployment, Model Training, Model Monitoring

C Model Deployment, Data Collection, Model Training, Model Monitoring

D Data Collection, Model Training, Model Deployment, Model Monitoring

💡 Hint

Think about what you need before training a model and what happens after deployment.

The ML lifecycle starts with collecting data, then training the model, deploying it, and finally monitoring its performance.

💻 Command Output

intermediate

Output of model evaluation metric calculation

Given the following Python code snippet that calculates accuracy, what is the output?

MLOps

from sklearn.metrics import accuracy_score
true_labels = [1, 0, 1, 1, 0]
predicted_labels = [1, 0, 0, 1, 0]
accuracy = accuracy_score(true_labels, predicted_labels)
print(f"Accuracy: {accuracy:.2f}")

A Accuracy: 1.00

B Accuracy: 0.60

C Accuracy: 0.80

D Accuracy: 0.40

💡 Hint

Count how many predictions match the true labels and divide by total.

4 out of 5 predictions are correct, so accuracy is 0.8.

🔀 Workflow

advanced

Correct sequence of steps for model deployment

Which option shows the correct sequence of steps to deploy a machine learning model in production?

A Deploy to cloud, Containerize model, Monitor model, Set up CI/CD pipeline

B Containerize model, Set up CI/CD pipeline, Deploy to cloud, Monitor model

C Monitor model, Deploy to cloud, Containerize model, Set up CI/CD pipeline

D Set up CI/CD pipeline, Deploy to cloud, Containerize model, Monitor model

💡 Hint

Think about preparing the model first, then automating deployment, then monitoring.

First, package the model (containerize), then automate deployment (CI/CD), deploy it, and finally monitor.

❓ Troubleshoot

advanced

Troubleshoot model monitoring alert

A deployed ML model suddenly shows a drop in accuracy. Which is the most likely cause?

A Data drift causing input data to differ from training data

B Model training completed successfully

C Model deployment pipeline is running smoothly

D Monitoring system is turned off

💡 Hint

Think about what can change in data after deployment.

Data drift means new data is different from training data, causing accuracy to drop.

✅ Best Practice

expert

Best practice for continuous integration in ML lifecycle

Which practice best supports continuous integration (CI) in the ML lifecycle?

A Automate testing of data preprocessing, model training, and deployment scripts

B Manually run model training after every code change

C Deploy models without version control to speed up delivery

D Ignore data validation to focus on faster model training

💡 Hint

CI means automating checks to catch problems early.

Automating tests for all ML pipeline parts ensures reliable integration and faster feedback.