questions
92 questions across 5 topics
92 questions
- Walk me through how transformers workML Breadthmedium
- Explain how RNNs work and where they fit todayML Breadthmedium
- Walk me through how modern object detection worksML Breadthmedium
- Explain logistic regression and why it's still used in industryML Breadtheasy
- Walk me through gradient descent variantsML Breadthmedium
- Walk me through backpropagationML Breadthmedium
- Walk me through how CNNs workML Breadthmedium
- What is regularization and how do the different techniques compare?ML Breadthmedium
- Explain ensemble methods — random forests and gradient boostingML Breadthmedium
- What are embeddings and how are they learned?ML Breadthmedium
- Explain cross-entropy loss — binary vs categorical, and why it works for classificationML Breadthmedium
- Compare MSE, MAE, and Huber loss — when would you use each?ML Breadthmedium
- What is focal loss and when would you use it over cross-entropy?ML Breadthhard
- What is KL divergence and how does it relate to cross-entropy?ML Breadthmedium
- Explain contrastive and triplet loss — how do they train embedding spaces?ML Breadthhard
- Walk me through feature engineering for tabular dataML Breadthmedium
- How do you encode categorical features? Compare one-hot, target encoding, and embeddings.ML Breadthmedium
- How do you handle class imbalance in ML?ML Breadthmedium
- How do you approach feature selection?ML Breadthmedium
- What is data leakage and how do you prevent it?ML Breadthmedium
- Walk me through classification evaluation metrics — precision, recall, F1, AUCML Breadthmedium
- Explain the bias-variance tradeoffML Breadthmedium
- What is model calibration and how do you fix a miscalibrated model?ML Breadthhard
- Compare batch norm, layer norm, and group norm — when do you use each?ML Breadthhard
- Compare SGD, Adam, and AdamW — when would you use each?ML Breadthmedium
- Explain attention mechanisms beyond transformers — types and trade-offsML Breadthhard
- Explain learning rate scheduling strategies and how to pick oneML Breadthmedium
- How do decision trees split? Walk me through Gini impurity vs information gain.ML Breadthmedium
- Explain transfer learning — when to freeze vs fine-tune, and common failure modesML Breadthmedium
- Explain NDCG — why use it over AUC for ranking?ML Breadthmedium
- What is train-serve skew and how do you prevent it?ML Breadthhard
- What is position bias in ranking and how do you correct for it?ML Breadthhard
- Walk me through two-tower models for retrieval — architecture, tradeoffs, and limitsML Breadthhard
- GBDT vs neural networks for ranking — when do you pick each?ML Breadthmedium
- How do you handle the cold start problem in recommendation systems?ML Breadthhard
- Your offline metrics improved but A/B test shows nothing — how do you debug?ML Breadthhard
- How does multi-task learning work for ranking — and what are the failure modes?ML Breadthhard
- Which features are safe to use when predicting future events — temporal leakage in practiceML Breadthhard
- Your CTR model's calibration degrades months after launch — what causes it and how do you fix it?ML Breadthhard
- Your model accuracy dropped 8% over 2 weeks — walk through your debugging processML Breadthhard
- Why don't huge overparameterized models always overfit?ML Breadthhard
- When is a large foundation model unjustified? Simple models vs LLMs.ML Breadthhard
- Why do residual connections make deep networks trainable?ML Breadthhard
- How do you initialize deep networks, and why does it matter?ML Breadthmedium
- What is a KV cache, and why does it matter for LLM inference?ML Breadthhard
- How do long-context transformers work, and where do they fail?ML Breadthhard
- How do mixture-of-experts models scale, and what can go wrong?ML Breadthhard
- When do you use a cross-encoder reranker instead of a bi-encoder?ML Breadthhard
- Explain SVMs and kernels from an interviewer's point of viewML Breadthmedium
- When would you use monotonic constraints in gradient boosted trees?ML Breadthhard
- What is gradient clipping, and when should you use it?ML Breadthmedium
- How does mixed precision training work, and why can it be unstable?ML Breadthhard
- Compare pointwise, pairwise, and listwise ranking lossesML Breadthhard
- How do label smoothing and knowledge distillation change cross-entropy training?ML Breadthhard
- How do you handle missing data without introducing bias?ML Breadthhard
- When do feature crosses matter, and how do you control their complexity?ML Breadthmedium
- How do delayed labels and censoring affect model training?ML Breadthhard
- What are point-in-time joins, and why do they prevent leakage?ML Breadthhard
- How do you handle domain shift and out-of-distribution inputs?ML Breadthhard
- How does data augmentation improve generalization, and when can it hurt?ML Breadthmedium
- How do you evaluate a new ranker using logs from an old policy?ML Breadthhard
- How do you put confidence intervals on ML metrics?ML Breadthhard
- How do canary and shadow deployments reduce ML launch risk?ML Breadthhard
- When should an ML system use human review instead of full automation?ML Breadthhard
- What is DSPy and when should you use it over manual prompting?ML Breadthhard
- How do you run A/B tests for ML models in production?ML Breadthhard
- When do you use Huber loss instead of MSE or MAE?ML Breadthmedium
- How would you design a recommendation system for Reddit?ML System Designhard
- How would you design a feature store for a large ML platform?ML System Designhard
- How would you design Amazon's product search ranking system?ML System Designhard
- How would you design YouTube's harmful content detection system?ML System Designhard
- How would you implement CI/CD for ML pipelines?MLOpsmedium
- How do you monitor a deployed model and detect drift?MLOpsmedium
- How would you deploy a new model version safely?MLOpsmedium
- What is a feature store and when do you actually need one?MLOpsmedium
- When do you retrain, and how do you make it reliable enough to automate?MLOpsmedium
- How do you track experiments and manage model versions?MLOpsmedium
- How do you serve ML models at low latency and high throughput?MLOpshard
- How do you run an A/B test to evaluate a new ML model?MLOpsmedium
- How do you build reliable data pipelines for ML?MLOpsmedium
- How do you make a model smaller and faster without losing much accuracy?MLOpshard
- Implement softmax and cross-entropy loss from scratchML Codingmedium
- Implement scaled dot-product attention from scratchML Codinghard
- Implement k-nearest neighbors with efficient distance computationML Codingmedium
- Implement gradient descent for linear regression from scratchML Codingeasy
- Implement a decision tree classifier from scratchML Codinghard
- Implement Word2Vec (skip-gram) training from scratchML Codinghard
- Walk me through how you'd design and analyze an A/B testStatistics & Probabilitymedium
- Explain Bayesian vs. frequentist thinking — when does each matter in ML?Statistics & Probabilitymedium
- Explain p-values, Type I/II errors, and statistical power to a non-statisticianStatistics & Probabilityeasy
- Which probability distributions matter most in ML, and when do you use each?Statistics & Probabilitymedium
- Your dataset has sampling bias — how do you detect and correct it?Statistics & Probabilityhard