AI Interview Questions

Introduction

A consolidated bank of common AI engineer interview questions across fundamentals, prompting, RAG, agents, deployment and architecture.

Understanding the topic

Core concepts to understand:

• What is an LLM and how does it work?
• Difference between fine-tuning and RAG?
• Walk through your last AI project.
• How do you prevent hallucinations?
• How do you evaluate a prompt?
• What is prompt injection and how do you defend?
• Compare GPT-4 vs Claude vs Llama 3.
• Design a multi-tenant RAG system.
• How do you deploy and monitor an AI feature?
• Trade-offs between agents vs simple workflows.
• How do you reduce LLM cost?
• Walk through chain-of-thought and ReAct.
• What is OWASP LLM Top 10?
• How would you build a copilot for our product?

Syntax reference

Visual workflow / architecture:

Levels of AI questions
─────────────────────────────────────
Beginner    → definitions, basic flows
Mid-level   → architecture, evals, costs
Senior+     → trade-offs, scaling, governance, system design

Real-world use

Pulled from real loops at Anthropic, OpenAI, Cohere, Cursor, Perplexity, Google DeepMind, Meta AI.

Best practices

• Practise out loud.
• Always discuss trade-offs.
• Bring real metrics from your projects.

Hands-on exercise

Interview preparation — practice these questions:

• Q1. Pick three questions and answer aloud.
• Q2. Where are you weakest?
• Q3. Build a 30-day study plan to fix it.