Katkat
Back

Explain core ML concepts and lifecycle

Capital OneMachine Learning Engineer · Onsite
MediumMachine Learning EngineerOnsiteMachine Learning

Last updated

This question evaluates understanding of core machine learning concepts and lifecycle competencies, including bias–variance tradeoff, parameters versus hyperparameters, batch versus real-time inference, strategies for training at scale, feature engineering, and end-to-end model development.

You are interviewing for an ML Engineer role. Answer the following (conceptually; no code required):

1) Bias–variance tradeoff

  • What are bias and variance?
  • How do they relate to underfitting vs. overfitting?
  • Name practical levers to move along the tradeoff.

2) Parameters vs. hyperparameters

  • Define each and give common examples.
  • Who/what “sets” each (training vs. tuning)?

3) Batch inference vs. real-time inference

  • Contrast latency, throughput, cost, and typical use cases.
  • What changes in feature computation and serving architecture?

4) Training with very large datasets

  • If training data is extremely large (doesn’t fit on a single machine), what strategies would you use?

5) Feature engineering

  • What is feature engineering?
  • Why does it matter even with modern models?

6) Walk through an end-to-end ML lifecycle

  • Describe an end-to-end ML lifecycle from problem framing to deployment and iteration.
  • Then walk through one end-to-end project you personally delivered (problem, data, modeling, evaluation, deployment, monitoring, iteration).