ML Interview Prep

Home Interview Journey Practice Library Mock Interview Progress

Question Bank

Practice algorithm and coding problems to ace your technical interviews.

ML/AI Interview Prep Platform - Your path to ML/AI Engineer roles

Built for the ML/AI community

44 real interview questions collected so far

Questions I personally encountered in interviews — more coming as I interview! These are shown first with a red background.

Recommended Next

0 of 113 completed. Here's a great next challenge:

Climbing Stairs

dynamic-programming, mathLeetCode #70

Your Progress

0 of 113 completed

0%

Source

Difficulty

Status

Topic

Linear Regression Implementation

machine-learning, linear-regression, coding +1Real Interview

What is Regularization?

machine-learning, regularization, overfitting +1Real Interview

Explain LoRA and QLoRA. How do they work and when would you use them?

lora, qlora, fine-tuning +3

PyTorch Tensor Operations

PyTorch, Tensors, Basic Operations +1

Do you have direct experience improving RAG performance in real-world applications?

RAG, Retrieval, Embeddings +3

Tell me about your last ML project. What challenges did you face and how did you solve them?

langgraph, multi-agent, optimization +2

Python OpenAI LangChain RAG

RAG, LangChain, FAISS +3

PyTorch Statistics

PyTorch, Tensors, Reshape +2

What's the difference between Tools and MCP (Model Context Protocol)?

mcp, tools, llm-integration +2

Python OpenAI Haystack RAG Project

RAG, Haystack, BM25 +3

Have you implemented an MCP server? Walk me through how you would build one.

mcp, fastmcp, server +2

What is the A2A (Agent-to-Agent) protocol? How does it compare to MCP?

a2a, google, multi-agent +2

Rate Limiter Implementation

Rate Limiter, System Design, TypeScript +2

Tell me about your LangGraph experience. How do you ensure correct results and visualize agent connections?

langgraph, multi-agent, validation +2

How do you deploy ML models in production? Walk me through your approach.

deployment, fastapi, docker +2

When would you fine-tune an LLM vs using prompting?

llm-engineering, fine-tuning, must-know

What model parameters do you adjust when using LLM APIs? What API types are available?

openai, api, parameters +2

Explain function calling in LLMs. How does the model decide when to use tools?

llm-engineering, function-calling, tools

What causes LLM hallucinations? How do you detect and prevent them?

llm-engineering, hallucination, must-know

Compare GPT-4o vs GPT-4o-mini vs other models. When do you use each?

model-selection, gpt-4o, cost-optimization +1

What vector databases have you used? Explain the different similarity metrics.

vector-db, pinecone, chromadb +3

How do you evaluate LLM outputs? What metrics matter?

llm-engineering, evaluation, metrics

How do you manage context windows in LLM applications?

llm-engineering, context-window, optimization

What agent frameworks have you used? Compare LangChain, LangGraph, and alternatives.

LLM, Production ML, Real Interview

You don't have customer data yet. How do you build and validate an ML system?

data-strategy, synthetic-data, few-shot +2

How do you evaluate RAG system accuracy? Have you used RAGAS?

ragas, evaluation, rag +3

Explain RAG (Retrieval-Augmented Generation). When and how would you use it?

llm-engineering, rag, retrieval +1

When is RAG not helpful? What are the limitations of Retrieval-Augmented Generation?

LLM, Production ML, Real Interview

How do you improve model accuracy in production ML systems?

LLM, Production ML, Real Interview

Compare GPT-4, Claude, and open-source LLMs. When would you use each?

llm-engineering, model-selection, cost-optimization

How do you reduce latency in LLM applications?

LLM, Production ML, Real Interview

How do you ensure LLMs output valid structured data (JSON, specific formats)?

llm-engineering, structured-outputs, validation

AI Orchestration: How do you coordinate multiple AI agents in workflows?

llm-engineering, multi-agent, orchestration +1

How do you quickly test and compare different LLM models?

LLM, Production ML, Real Interview

Explain PCA (Principal Component Analysis). When and how do you use it?

dimensionality-reduction, feature-engineering, unsupervised +1

How do you choose the number of epochs? Explain early stopping and gradient descent optimizers.

training, optimization, epochs +3

When would you use fine-tuning vs RAG vs prompt engineering? How do you decide?

llm-engineering, fine-tuning, rag +2

What performance/load testing tools have you used? How do you create realistic load tests?

load-testing, performance, locust +2

What do you do when traffic is too high? How do you handle traffic spikes?

scaling, auto-scaling, rate-limiting +2

How do you identify and fix slow endpoints?

performance, optimization, profiling +3

What challenges have you faced in full-stack development and deployment? How did you solve them?

deployment, full-stack, docker +4

Filter Words by Letters

strings, sets, filtering +3Tech Company

Factorize String Extremities

strings, greedy, string-matching +2Tech Company

Sequence Join Point

sequences, hash-set, math +3Tech Company

Regularization: L1, L2, and Beyond

machine-learning, regularization, overfitting +1Generated

Climbing Stairs

dynamic-programming, math, memoizationLeetCode #70

Number of Islands

graph, dfs, bfs +1LeetCode #200

Contains Duplicate

array, hash-tableLeetCode #217

Valid Anagram

hash-table, string, sortingLeetCode #242

Reverse Linked List

linked-list, recursionLeetCode #206

Valid Parentheses

stack, stringLeetCode #20

Merge Two Sorted Lists

linked-list, recursionLeetCode #21

Invert Binary Tree

tree, dfs, recursionLeetCode #226

Binary Tree Level Order Traversal

tree, bfs, queueLeetCode #102

Min Stack

stack, designLeetCode #155

Kth Largest Element in an Array

heap, quickselect, sortingLeetCode #215

Clone Graph

graph, dfs, bfs +1LeetCode #133

Group Anagrams

hash-table, string, sortingLeetCode #49

Top K Frequent Elements

array, hash-table, heap +1LeetCode #347

Merge K Sorted Lists

linked-list, heap, divide-and-conquerLeetCode #23

Same Tree

tree, dfs, recursionLeetCode #100

Two Sum

array, hash-tableLeetCode #1

Maximum Depth of Binary Tree

tree, dfs, recursionLeetCode #104

Linked List Cycle

linked-list, two-pointersLeetCode #141

Coin Change

dynamic-programming, array, bfsLeetCode #322

System Design: AI Customer Support with AWS Bedrock

AWS, Bedrock, System Design +5

System Design: MLOps Pipeline with MLflow & Kubeflow

Kubernetes, Kubeflow, MLflow +4

System Design: LangGraph Log Analysis & Slack Notification System

LangGraph, Multi-Agent, System Design +5

What is overfitting? How do you prevent it?

fundamental, regularization, must-know

What is Model Context Protocol (MCP) and how would you use it to extend LLM capabilities?

MCP, LLM, Tool Use +3

Explain the bias-variance tradeoff.

fundamental, theory, must-know

Explain backpropagation. How does it work?

deep-learning, optimization, must-know

Explain precision, recall, and F1 score. When to optimize for which?

evaluation, metrics, must-know

What is the attention mechanism in transformers?

nlp, transformers, deep-learning +1

Compare zero-shot, few-shot, and chain-of-thought prompting. When would you use each?

llm-engineering, prompting, must-know

How do you evaluate LLM outputs? What metrics and methods do you use?

llm-evaluation, bleu, rouge +2

How do you detect and prevent hallucinations in LLM applications?

hallucination, groundedness, factuality +2

How do you run A/B tests for LLM-powered features?

a/b-testing, experimentation, production-ml +1

How do you evaluate and test an LLM-based system before deploying to production? What metrics do you track?

llm-testing, production-ml, rag-evaluation +3

How do you handle model versioning and rollback in production? What happens if a new model performs worse than expected?

model-versioning, rollback, mlops +3

Walk me through building a production fine-tuning pipeline from start to finish. What are the key steps?

fine-tuning, production-ml, mlops +5

How do you prepare training data for fine-tuning? How much data do you need and how do you ensure quality?

llm-engineering, fine-tuning, data-preparation +1

What is a knowledge graph and how does it differ from a traditional relational database?

knowledge-graph, neo4j, graph-database +4

What is an ontology in the context of Knowledge Graphs? Why is it important?

knowledge-graph, ontology, rdf +4

How can Knowledge Graphs help reduce hallucinations in LLM applications?

knowledge-graph, hallucination, rag +4

How do you approach feature engineering for ML models? Walk me through your process.

feature-engineering, data-science, preprocessing +1

How do you ensure data quality in production ML pipelines? What tools and practices do you use?

data-quality, great-expectations, pandera +2

Explain your approach to data aggregation for analytics and ML features. How do you handle different time windows?

aggregation, sql, window-functions +1

How do you validate data at different stages of a pipeline? What tools do you use?

data-validation, pandera, great-expectations +1

How do you optimize slow SQL queries for large datasets? Walk me through your debugging process.

sql, optimization, indexing +1

Compare ETL vs ELT approaches. When would you use each?

etl, elt, dbt +2

How do you handle missing data in ML pipelines? What are the different strategies?

missing-data, imputation, preprocessing +1

How do you detect and handle data drift in production ML systems?

data-drift, monitoring, mlops +1

You mention 'Python (Advanced)' on your resume. Explain Python's GIL and how it affects multi-threading in ML workloads.

python, concurrency, performance +1

Explain Python's type hints and how you use them in production ML code. Why are they important?

python, type-hints, code-quality +1

Walk me through your code structure for a production ML API. What design patterns do you use?

python, architecture, design-patterns +1

Your ML interview website is built with TypeScript. Why TypeScript over JavaScript for this project?

typescript, frontend, type-safety +1

You mention React experience from Android/React Native. How did you apply that to building web interfaces for your ML projects?

react, frontend, gradio +1

Why did you choose Next.js for your ML interview website instead of plain React?

next.js, ssg, ssr +2

You've built 4 production FastAPI services. What are your REST API design principles?

rest-api, fastapi, api-design +1

You use pgvector for semantic search in your RAG chatbot. How does it work and why PostgreSQL over a dedicated vector DB?

postgresql, pgvector, vector-search +1

You mention 35% cache hit rate with Redis. Walk me through your caching strategy.

redis, caching, optimization +1

You fine-tuned BERT for sentiment analysis (89% F1). Explain the fine-tuning process step-by-step.

bert, nlp, fine-tuning +1

You used ResNet50 for image classification (94% accuracy). Why ResNet over other architectures?

resnet, computer-vision, transfer-learning

You implemented Grad-CAM for model interpretability. How does it work and why is it useful?

grad-cam, interpretability, explainability +1

You mention AWS (S3, EC2, SageMaker). How have you used these services in your ML projects?

aws, s3, ec2 +2

Implement K-Means Clustering

clustering, unsupervised, classic-ml +2

Implement Linear Regression with Gradient Descent

regression, gradient-descent, optimization +2

Implement Softmax and Cross-Entropy Loss

neural-networks, classification, loss-functions +2

Golden Brick Road

bfs, graph, shortest-path +3

Fibonacci Number

dynamic-programming, recursion, easy +2

Valid Palindrome

two-pointers, string, easy +2

Reverse an Array

two-pointers, array, easy +2