amitshekhariitbhu/ai-engineering-interview-questions

What this project is

A free, open-source study guide for people preparing for AI engineering job interviews. It's a long list of interview questions with answers, organized by topic.

What it contains

The questions cover 15 topic areas, including:

• LLM fundamentals — how large language models work under the hood (transformers, attention, tokenization, sampling, KV cache, etc.)
• Prompt engineering — techniques for getting reliable output from LLMs (few-shot prompting, chain-of-thought, prompt injection defense, etc.)
• Retrieval-Augmented Generation (RAG) — building systems that retrieve relevant documents before generating answers
• AI agents and agentic systems — autonomous systems that reason and take actions
• Fine-tuning and model adaptation — customizing pre-trained models
• Vector databases and embeddings — storing and searching text as vectors
• AI system design — architecture for production AI systems
• LLMOps and production AI — deploying, monitoring, and maintaining LLMs in production
• Evaluation and testing — measuring model quality
• AI safety, ethics, and responsible AI — risks and safeguards
• Multimodal AI — models that handle images, audio, and text
• AI infrastructure and scalability — serving models at scale
• Coding and practical implementation — hands-on programming questions
• Behavioral and scenario-based questions — situational interview questions

Many questions include a link to a detailed answer on the maintainer's blog, YouTube channel, or social media. Some questions have no linked answer yet.

Who made it

Amit Shekhar, founder of Outcome School (an education platform for AI and machine learning). He maintains it and says it will be updated continuously with new questions.

Who it's for

People interviewing for roles such as AI Engineer, Gen AI Engineer, LLM Engineer, Agentic AI Engineer, AI Solutions Architect, MLOps Engineer, and similar positions.

Notable details

• Includes a "Must Know" section covering six core concepts: LLM, RAG, MCP, Agent, Fine-tuning, and Quantization.
• Answers are not all written inline — many point to external articles and videos.
• The project is actively maintained (the README states it will keep being updated).