System Design

Question 1

What is a data pipeline?

Answer 1

Automated process for collecting, processing, and storing data for ML models.

Question 2

What is model deployment?

Answer 2

Making an ML model available for predictions in a production environment.

Question 3

What is batch prediction?

Answer 3

Generating predictions for large datasets periodically.

Question 4

What is online prediction?

Answer 4

Serving predictions in real-time via APIs.

Question 5

What is model monitoring?

Answer 5

Tracking performance metrics post-deployment to detect drift.

Question 6

What is concept drift?

Answer 6

Change in the relationship between input and output variables over time.

Question 7

What is feature store?

Answer 7

Centralized repository for storing and serving ML features.

Question 8

What is canary deployment?

Answer 8

Rolling out model updates to a small subset of users before full deployment.

Question 9

What is A/B testing in ML?

Answer 9

Comparing performance of two models or variants to see which performs better.

Question 10

What is shadow deployment?

Answer 10

Deploying a model in parallel to production without affecting user-facing results.