Data Science & ML

MLOps & Deployment

MLflow, Docker, FastAPI, Streamlit, Prefect, feature stores, data pipelines, monitoring, ML business metrics, cloud deployment

24 interview questionsยท
Senior
1

What is the primary role of MLflow in an MLOps workflow?

Answer

MLflow is an open-source platform that manages the complete ML model lifecycle: experiment tracking (metrics, parameters, artifacts), model packaging, centralized registry and deployment. This enables experiment reproducibility and standardized model versioning.

2

Which command is used to log a parameter in MLflow?

Answer

The mlflow.log_param function records a hyperparameter (learning rate, epochs, batch size) associated with a run. These parameters are then visible in the MLflow UI and allow comparison of different training configurations.

3

What is the difference between mlflow.log_metric and mlflow.log_param?

Answer

log_param records fixed values defined before training (hyperparameters like learning_rate, epochs), while log_metric records values that change during or after training (accuracy, loss). Metrics can be logged multiple times with different steps to create curves.

4

What is the main advantage of using Docker to deploy an ML model?

5

Why use a multi-stage Dockerfile for an ML application?

+21 interview questions

Master Data Science & ML for your next interview

Access all questions, flashcards, technical tests, code review exercises and interview simulators.

Start for free