Personalized prep

Stop memorizing.
Start understanding.

Your personalized coach for AI, ML and GenAI interviews. Add your resume, the job description, and your interview date, and crackAI.dev builds a prep plan tuned to what the role actually tests.

Your prep plan

Google · ML Engineer (L4) · paced to your date

Interview in 3 days
T-10 days

Study core ML topics, plus a coding lab to warm up

T-7 days

ML depth: RAG, embeddings & transformers topics

T-3 daysUp next

Design a RAG system, run a mock, prep behavioral stories

build_rag.py
12345678910

def answer(query, k=5):

# 1. embed + retrieve top-k chunks

q = embed(query)

hits = index.search(q, k=k)

ctx = rerank(query, hits)[:3]

 

# 2. ground the LLM in retrieved context

prompt = build_prompt(query, ctx)

return llm.generate(prompt)

All tests passed (12ms)
0+Interview Questions
0+Core Papers Analyzed
0Hands-on Code Labs
0System Design Labs

How crackAI.dev preps you

A personalized interview coach for AI, ML and GenAI engineers. Three inputs in, a plan and the practice to back it out, focused on what the role actually tests.

01

Add resume, JD and date

Drop in your resume, paste the job description (or a link), and set your interview date. crackAI.dev reads the role, the stack, and what it will test.

02

Get a plan tuned to you

A prep plan built for your target role and timeline: what to learn, what to revise, and what to skip, pacing you toward your date.

03

Run real mock interviews

Mock interviews built from your target role and your own projects, graded with feedback, so practice mirrors the loop you will actually face.

Preparation goes deep and wide

From concept to working system. Take RAG, end to end: from “What is RAG?” to “Why RAG over fine-tuning?” to designing a system to building one in an AI pair-programmed coding lab.

UnderstandWhat is RAG?
ReasonWhy RAG over fine-tuning?
DesignDesign a RAG system
BuildBuild it in a code lab

It works the same way across the system design labs these roles depend on, from agents to search and recommenders, at beginner to advanced levels.

Everything the role tests, in one place

The practice surfaces AI, ML and GenAI interviews are built on: design, code, fundamentals, and the literature behind them.

◣

System Design Labs

Design the architectures these roles depend on (agents, RAG, search, recommenders) with multi-level interview progression. Start simple and scale up through tougher requirements, just like a real loop.

Explore design labs
⌘

CodeLabs

AI pair-programmed coding labs. Build the systems you just designed, from RAG pipelines to training loops, with a built-in assistant to debug, optimize, and ship alongside you.

Explore coding labs
?

Interview Questions

Curated questions spanning ML fundamentals, deep learning, transformers, GenAI and RAG, plus behavioral, with model answers and company tags.

Browse interview questions
❖

Papers & Research

Study the foundational and frontier papers behind modern AI with structured summaries, key takeaways, and interview-relevant insights.

Browse papers
✉

Discussions

Share interview experiences, exchange study tips, post job referrals, and help each other prepare. A community-driven forum with upvotes, anonymous posting, and company tags.

Join the discussion

Who's behind it

Built by AI engineers from Microsoft, AMD, IIT Bombay and IIIT Hyderabad, with published AI research.

MicrosoftAMDIIT BombayIIIT Hyderabad

Just want to prepare?

Set a target date and start building toward it. No interview on the calendar required.

No date yet? Set a target and start building →

Set a target date→Browse practice

Built on the literature you'll be asked about

Attention Is All You Need2017BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding2018Language Models are Few-Shot Learners2020Scaling Laws for Neural Language Models2020Training Language Models to Follow Instructions with Human Feedback2022Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks2020Chain-of-Thought Prompting Elicits Reasoning in Large Language Models2022LoRA: Low-Rank Adaptation of Large Language Models2021Learning Transferable Visual Models From Natural Language Supervision2021High-Resolution Image Synthesis with Latent Diffusion Models2022Deep Residual Learning for Image Recognition2015Adam: A Method for Stochastic Optimization2014Generative Adversarial Nets2014Training Compute-Optimal Large Language Models2022
Attention Is All You Need2017BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding2018Language Models are Few-Shot Learners2020Scaling Laws for Neural Language Models2020Training Language Models to Follow Instructions with Human Feedback2022Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks2020Chain-of-Thought Prompting Elicits Reasoning in Large Language Models2022LoRA: Low-Rank Adaptation of Large Language Models2021Learning Transferable Visual Models From Natural Language Supervision2021High-Resolution Image Synthesis with Latent Diffusion Models2022Deep Residual Learning for Image Recognition2015Adam: A Method for Stochastic Optimization2014Generative Adversarial Nets2014Training Compute-Optimal Large Language Models2022
Attention Is All You Need2017BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding2018Language Models are Few-Shot Learners2020Scaling Laws for Neural Language Models2020Training Language Models to Follow Instructions with Human Feedback2022Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks2020Chain-of-Thought Prompting Elicits Reasoning in Large Language Models2022LoRA: Low-Rank Adaptation of Large Language Models2021Learning Transferable Visual Models From Natural Language Supervision2021High-Resolution Image Synthesis with Latent Diffusion Models2022Deep Residual Learning for Image Recognition2015Adam: A Method for Stochastic Optimization2014Generative Adversarial Nets2014Training Compute-Optimal Large Language Models2022
Attention Is All You Need2017BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding2018Language Models are Few-Shot Learners2020Scaling Laws for Neural Language Models2020Training Language Models to Follow Instructions with Human Feedback2022Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks2020Chain-of-Thought Prompting Elicits Reasoning in Large Language Models2022LoRA: Low-Rank Adaptation of Large Language Models2021Learning Transferable Visual Models From Natural Language Supervision2021High-Resolution Image Synthesis with Latent Diffusion Models2022Deep Residual Learning for Image Recognition2015Adam: A Method for Stochastic Optimization2014Generative Adversarial Nets2014Training Compute-Optimal Large Language Models2022