Interview Prep

Technical topics, coding challenges, behavioral questions and system design — everything to land your AI role.

Preparing for a machine learning interview means covering four surfaces at once: technical ML topics, coding, behavioral and system design. This page organizes all of it into ten tabs — a 200+ question bank with answer key points and follow-ups, real questions tagged by company and difficulty, and curated prep resources — so AI interview questions stop being a guessing game.

Two tabs go beyond the candidate's view. The interviewer's view shows what real interviewers at big tech, startups and research labs reward and flag, and the hiring-process tab maps timelines, rounds and pass rates per company. A dedicated China section covers the bǐshì written test that US processes usually skip. Start with the technical topics tab and work across.

Explore: ML interview questions · ML interviews at top companies · How to prepare

Machine Learning Fundamentals

Supervised Learning
RegressionClassificationOverfittingRegularization
  • Explain bias-variance tradeoff
  • What is regularization and why is it important?
  • Difference between L1 and L2 regularization
  • How do you handle imbalanced datasets?
Model Evaluation
Cross-validationMetricsROC-AUCPrecision-Recall
  • When to use precision vs recall?
  • Explain cross-validation and its types
  • How to evaluate regression models?
  • What is the difference between accuracy and F1-score?
Feature Engineering
ScalingEncodingFeature SelectionDimensionality Reduction
  • How do you handle categorical variables?
  • When to normalize vs standardize?
  • Explain PCA and when to use it
  • How do you detect and handle outliers?

Deep Learning

Neural Networks Basics
BackpropagationActivation FunctionsOptimizersLoss Functions
  • Explain backpropagation
  • Why use ReLU over sigmoid?
  • What is vanishing gradient problem?
  • Compare SGD, Adam, and RMSprop
CNNs
ConvolutionPoolingResNetTransfer Learning
  • How does convolution work?
  • Explain pooling and its types
  • What are skip connections?
  • When to use transfer learning?
Transformers & LLMs
AttentionSelf-AttentionBERTGPT
  • Explain self-attention mechanism
  • What are positional encodings?
  • Difference between BERT and GPT
  • How does multi-head attention work?

LLM Applications

Prompt Engineering
Few-shot learningChain-of-thoughtSystem prompts
  • Best practices for prompt engineering
  • How to reduce hallucinations?
  • Explain few-shot prompting
  • When to use chain-of-thought?
RAG (Retrieval-Augmented Generation)
Vector searchEmbeddingsChunkingReranking
  • How does RAG work?
  • Explain vector embeddings
  • What is semantic search?
  • How to improve RAG accuracy?
Fine-tuning
LoRAPEFTRLHFInstruction tuning
  • When to fine-tune vs prompt engineer?
  • Explain LoRA
  • What is RLHF?
  • How to prepare fine-tuning data?

MLOps & Production

Model Deployment
APIsContainerizationServingMonitoring
  • How to deploy ML models?
  • What is model serving?
  • Explain A/B testing for models
  • How to monitor model performance?
Scalability
Distributed trainingInference optimizationCaching
  • How to scale ML systems?
  • Explain data parallelism vs model parallelism
  • How to optimize inference speed?
  • What is model quantization?

FAQ

Common questions, answered with the data on this page.

What questions are asked in a machine learning interview?

Expect a mix of ML theory, coding, behavioral and system design. The question bank above holds 200+ curated questions with answer key points and follow-ups, grouped by category and difficulty, and the company-questions tab shows real questions asked at specific companies. Our machine learning interview questions guide adds full sample answers.

What are ML interviews at top companies like?

Multi-round and staged: the hiring-process tab maps each company's timeline, round sequence and pass rates, while the interviewer's-view tab shows what interviewers at big tech, startups and research labs count as top qualities and red flags. Our guide to ML interviews at top companies walks through a full loop.

How should I prepare for an ML engineer interview?

Work the tabs left to right: shore up technical topics, drill the coding challenges, rehearse behavioral stories, then practice the system-design problems. The resources tab lists prep books, courses and practice platforms for each stage. A week-by-week plan is in our ML engineer interview preparation guide.

How are AI interviews in China different?

The China interview tab covers the key difference: a written coding test (bǐshì) that US processes usually skip, plus Chinese-style algorithm and AI questions, behavioral rounds and offer negotiation. If you are interviewing in both markets, prepare for the written test separately — it has its own format and timing.