AI/ML Engineer Interview Preparation
120 deep-dive questions across 9 tracks — from linear algebra and classical ML through transformers, LLM fine-tuning, RAG, agents, MLOps and ML system design. Every answer includes the derivation, the trade-offs, the mistakes to avoid and the follow-up questions interviewers actually ask.
- Lessons
- 24
- Deep-dive questions
- 120
- Quiz questions
- 120
- Hands-on challenges
- 24
Progress is saved in this browser only — no account, no tracking, nothing leaves your device.
Nine tracks, one path
Ordered so each track builds on the last. Skip ahead freely — every lesson stands on its own.
Math & Statistics Foundations
3 lessons · 15 questions
The linear algebra, probability and statistics that every ML screen quietly assumes you already know.
Start trackClassical Machine Learning
5 lessons · 25 questions
Regression, trees, ensembles, feature engineering and the evaluation metrics interviewers grill you on.
Start trackDeep Learning
3 lessons · 15 questions
Backprop, optimisers, normalisation, CNNs and the transformer maths you must be able to derive on a whiteboard.
Start trackLLMs & Applied NLP
3 lessons · 15 questions
Tokenisation, pre-training, LoRA/QLoRA, RLHF vs DPO, decoding, quantisation and LLM evaluation.
Start trackRAG & AI Agents
2 lessons · 10 questions
Retrieval pipelines, vector search, reranking, agent loops, tool calling and production guardrails.
Start trackMLOps & Production ML
2 lessons · 10 questions
Serving patterns, drift monitoring, CI/CD for models, cost control and inference optimisation.
Start trackML System Design
2 lessons · 10 questions
A repeatable 8-step framework plus worked case studies for recommenders, ranking, fraud and LLM products.
Start trackML Coding Rounds
2 lessons · 10 questions
NumPy/pandas fluency, algorithms from scratch and idiomatic PyTorch under interview time pressure.
Start trackBehavioural & Strategy
2 lessons · 10 questions
Project deep-dives, STAR stories, levelling expectations, take-homes and offer negotiation.
Start trackBuilt the way interviews actually go
Interviewers do not stop at your first answer. Neither does this course.
Answers with the reasoning, not just the fact
Every question has a 30-second spoken answer, a full derivation, the key points to say, the mistakes that sink candidates, and the follow-ups you will actually get.
Runnable code with real output
Over 200 verified code examples in NumPy, PyTorch, pandas and SQL — with the printed output, so you can see why the answer is what it is.
Mock interview simulator
Timed random questions filtered by track and difficulty. Answer out loud, reveal the model answer, self-grade, and get a review list of your weak areas.
A 12-week plan that ends in offers
Four phases from maths foundations to offer negotiation, with a weekly goal, the lessons to cover, and a concrete deliverable you can talk about in a deep-dive.
The full curriculum
24 lessons across 9 tracks. Search by topic, company, or the exact question you were asked.
Math & Statistics Foundations
The linear algebra, probability and statistics that every ML screen quietly assumes you already know.
Linear Algebra for Machine Learning Interviews
Every embedding, attention head and gradient update is linear algebra. Master the five questions interviewers actually ask about vectors, matrix decompositions and derivatives.
Probability & Bayesian Thinking for ML Interviews
The probability questions that separate candidates: base rates, why MLE is just cross-entropy in disguise, MAP as regularisation, and the distribution choices baked into every loss function.
Statistics & A/B Testing for ML Engineers
Shipping a model means running an experiment. Learn how to size it, read it honestly, avoid peeking, and explain why offline AUC gains vanish online.
Classical Machine Learning
Regression, trees, ensembles, feature engineering and the evaluation metrics interviewers grill you on.
Machine Learning Fundamentals: The Screening Round
The five questions that open almost every ML interview. Answer them with diagnostics and trade-offs instead of definitions and you clear the screen in ten minutes.
Linear & Logistic Regression Deep Dive
Interviewers still open with linear models because they expose whether you understand optimisation, probability and diagnostics — or just call .fit().
Decision Trees, Random Forests & Gradient Boosting
Gradient boosting still wins most tabular problems in production. Know how it works internally, why it beats random forests, and how to explain its predictions.
Evaluation Metrics: Choosing and Defending the Right One
Picking the metric is a product decision disguised as a technical one. Learn to derive the right metric from the cost of each error type — and to defend it.
Feature Engineering, Imbalanced Data & Unsupervised Learning
Features still beat models on tabular problems. Learn the engineering patterns that win, the honest way to handle imbalance, and how to evaluate unsupervised models with no labels.
Deep Learning
Backprop, optimisers, normalisation, CNNs and the transformer maths you must be able to derive on a whiteboard.
Neural Networks & Backpropagation From First Principles
Derive backprop, explain why ReLU replaced sigmoid, and know exactly which initialisation to use and why — the three things every deep-learning interview verifies.
Optimisers, Schedules & Regularisation That Actually Work
The practical half of deep learning: how to pick an optimiser, set a learning rate, stop a run from diverging, and regularise without destroying capacity.
Transformers & Attention: The Full Derivation
The most-asked architecture question in AI interviews. Derive scaled dot-product attention, explain the √d, and know why grouped-query attention exists.
LLMs & Applied NLP
Tokenisation, pre-training, LoRA/QLoRA, RLHF vs DPO, decoding, quantisation and LLM evaluation.
LLM Fundamentals: Tokenisation, Pre-training & Scaling Laws
Everything downstream of the tokenizer depends on it. Understand BPE, the pre-training objectives, and the scaling laws that decide how big a model should be.
Fine-tuning, LoRA/QLoRA and Preference Optimisation
When to fine-tune at all, how LoRA actually works, and the real difference between RLHF and DPO — with the numbers that justify each choice.
LLM Inference: Decoding, Quantisation & Serving Economics
The half of LLM engineering that decides whether your product is profitable: decoding strategies, quantisation trade-offs, and how to get 10× throughput from the same GPU.
RAG & AI Agents
Retrieval pipelines, vector search, reranking, agent loops, tool calling and production guardrails.
RAG Architecture: Chunking, Vector Search & Hybrid Retrieval
RAG is the most-built AI system in industry and the most-failed. Learn the design decisions that actually determine whether it works: chunking, hybrid retrieval and index choice.
AI Agents, Tool Calling & Production Guardrails
Agents are easy to demo and hard to ship. Learn the loop, the failure modes, the guardrails, and when a deterministic pipeline beats an agent outright.
MLOps & Production ML
Serving patterns, drift monitoring, CI/CD for models, cost control and inference optimisation.
Model Deployment: Serving Patterns, Feature Stores & Rollouts
The gap between a notebook and a service. Serving patterns, the feature-store problem they solve, and how to ship a model without breaking production.
Monitoring, Drift Detection & Incident Response
Models decay silently. Learn what to monitor when labels arrive months late, how to distinguish drift types, and how to run an ML incident.
ML System Design
A repeatable 8-step framework plus worked case studies for recommenders, ranking, fraud and LLM products.
The ML System Design Framework (8 Steps)
A repeatable structure for the 45-minute ML system design round, plus the three moves that separate a hire from a no-hire in the first five minutes.
Search, Ranking & Feed Design Case Studies
Four more worked designs covering the surfaces most companies actually build: search, feeds, ads and moderation — each with the trap the interviewer is waiting for.
ML Coding Rounds
NumPy/pandas fluency, algorithms from scratch and idiomatic PyTorch under interview time pressure.
NumPy, pandas & ML Algorithms From Scratch
The ML coding round is not LeetCode. It is 'implement k-means in 20 minutes' and 'vectorise this loop' — with an interviewer watching how you debug.
PyTorch Fluency & Debugging Under Interview Pressure
Write a correct training loop from memory, explain autograd precisely, and debug a broken model live — the three things a PyTorch round actually tests.
Behavioural & Strategy
Project deep-dives, STAR stories, levelling expectations, take-homes and offer negotiation.
Project Deep-Dives & Behavioural Rounds for ML Engineers
The round that decides your level. How to tell a project story that survives ten follow-up questions, and how to answer 'tell me about a failure' without hurting yourself.
Interview Strategy, Take-Homes & Negotiating the Offer
The meta-game: which companies test what, how to win a take-home, what to ask your interviewers, and how to negotiate without risking the offer.
Practice tools
Reading is not preparation. These turn the material into recall you can use under pressure.
Flashcard drill
All 120 questions as spaced-recall cards. Filter by track, flip with the space bar, grade yourself.
OpenMock interview
Timed random questions. Answer out loud, reveal the model answer, self-grade, get a weak-area review list.
OpenGlossary
70+ terms with definitions and — more usefully — the specific angle an interviewer will push on.
OpenFrequently asked questions
Is this AI/ML interview course really free?
Yes. Every lesson, quiz, flashcard and the mock interview simulator are completely free with no sign-up. Your progress is stored in your browser's local storage and never leaves your device.
Who is this course for?
Anyone preparing for Machine Learning Engineer, Applied Scientist, AI Engineer or ML Platform roles. It assumes you can code in Python and have seen machine learning before — it is preparation, not a first introduction. Levels range from foundational to expert.
Does it cover LLMs, RAG and AI agents?
Extensively. There are dedicated tracks on LLM fundamentals and scaling laws, fine-tuning with LoRA/QLoRA, RLHF versus DPO, inference optimisation and quantisation, RAG architecture with chunking and hybrid retrieval, and production AI agents including prompt-injection defence.
How long does it take to complete?
Roughly 8 hours of reading across 24 lessons, plus the hands-on challenges. The included 12-week plan schedules 8-12 hours per week; you can compress it to six weeks at 20+ hours per week.
How is this different from a question list?
Question lists give you an answer to memorise. This gives you the derivation, the trade-off, the diagnostic process and the follow-up questions — because interviewers probe two layers past the first answer, and that is where candidates fail.
Do I need a PhD for AI/ML engineering roles?
Not for ML Engineer, AI Engineer or ML Platform roles, which are the majority of the market. Research Scientist positions usually do expect one. The behavioural track covers how the different role types are interviewed and what each expects.
Start where you are weakest
Run a five-question mock interview right now. Whatever you blank on is your first lesson.
Curriculum
- Machine Learning Fundamentals: The Screening Round5 Q · 18 min
- Linear & Logistic Regression Deep Dive5 Q · 19 min
- Decision Trees, Random Forests & Gradient Boosting5 Q · 20 min
- Evaluation Metrics: Choosing and Defending the Right One5 Q · 18 min
- Feature Engineering, Imbalanced Data & Unsupervised Learning5 Q · 19 min