---
title: "ML interview questions"
description: "Real ML interview questions by difficulty — core ML, LLMs, system design, Python, and computer vision — with tags to drill weak spots."
canonical_url: "https://fanout.sh/ai/interviews"
md_url: "https://fanout.sh/ai/interviews.md"
access: "public"
---

# ML interview questions

Real ML interview questions by difficulty — core ML, LLMs, system design, Python, and computer vision — with tags to drill weak spots.

## Public interview questions

### Overfitting vs. Underfitting — how to detect and fix each?

- Difficulty: easy
- Category: Core ML
- Tags: Theory, Basics

### Precision vs. Recall — when to optimize for which?

- Difficulty: easy
- Category: Core ML
- Tags: Metrics, Evaluation

### Explain Gradient Descent and its variants (SGD, Adam, AdaGrad).

- Difficulty: medium
- Category: Core ML
- Tags: Optimization, Math

### How to detect Exploding vs. Vanishing Gradients?

- Difficulty: medium
- Category: Core ML
- Tags: Debugging, Deep Learning

### Explain Feature Engineering — key techniques and when to use them.

- Difficulty: medium
- Category: Core ML
- Tags: Data Engineering, Process

---
This representation contains public Fanout content only. Protected Pro lessons, account data, billing, checkout, and pricing are not included.

Browse the public content map: https://fanout.sh/sitemap.md