Machine Learning Engineer Interview Questions

Showing 121–150 of 202 results

• OpenAIHard·Machine Learning Engineer
  Repair a Broken Tiny Language Model, Then Make Decoding Incremental

  Find four planted bugs in a compact PyTorch decoder-only language model (some produce NaNs, some silently ruin learning), then add key/value caching so generation processes one new token per step.

  Machine LearningOnsitepytorchtransformersdebugging
• OpenAIHard·Machine Learning Engineer
  Train a classifier and analyze dataset

  This question evaluates proficiency in end-to-end supervised machine learning workflows—covering data validation, exploratory data analysis, preprocessing, model training and selection, evaluation metrics and calibration, fairness assessments, and reproducibility—and tests competencies in applied machine learning engineering, statistical reasoning, and model governance within the Machine Learning domain. It is commonly asked in technical interviews because it verifies practical implementation skills and judgment for deploying reliable classifiers, aligning metrics to business goals and diagnosing performance and fairness across slices, with emphasis on practical application informed by conceptual understanding.

  Machine LearningTechnical Screen
• OpenAIMedium·Machine Learning Engineer
  Vectorize One-Nearest-Neighbor and Express It as a Neural Forward Pass

  Tests vectorized distance computation in array libraries and the mathematical equivalence between nearest-prototype classification and an affine neural network layer.

  Machine LearningTechnical Screenvectorizationnearest-neighborslinear-algebra
• PPPDT PartnersMedium·Machine Learning Engineer
  Project Deep Dive: Low-Latency LSTM Inference at Scale and the LSTM vs Transformer Choice

  A project deep dive on a small LSTM model served for large-scale, latency-sensitive inference. The candidate must explain the deployment hardware, latency requirement and use case, how inference latency was reduced, why an LSTM was chosen over a Transformer, and whether that choice still holds today.

  Machine LearningTechnical Screen
• PerplexityMedium·Machine Learning Engineer
  Train and improve a scikit-learn binary classifier

  This problem tests your proficiency in the scikit-learn workflow, specifically focusing on data preprocessing, addressing class imbalance in imbalanced datasets, and selecting appropriate metrics beyond accuracy.

  Machine LearningTechnical Screenscikit-learnImbalanced DataModel Evaluation
• PinterestHard·Machine Learning Engineer
  Answer core ML fundamentals questions

  This question evaluates a candidate's mastery of core Machine Learning fundamentals, including classification metrics (precision and recall) with confusion-matrix interpretation, gradient concepts for optimization, ensemble learning tradeoffs, and the mechanics of a neural network forward pass.

  Machine LearningOnline Assessment
• PinterestHard·Machine Learning Engineer
  Bias-Variance Tradeoff and Vanishing Gradients in Feedforward Networks

  Tests theoretical and mathematical understanding of the bias-variance decomposition and the vanishing-gradient phenomenon during backpropagation in deep feedforward architectures.

  Machine LearningTechnical ScreenDeep LearningBackpropagationModel Generalization
• PinterestMedium·Machine Learning Engineer
  Explain bias–variance, overfitting, and vanishing gradients

  This question tests the conceptual understanding of the bias-variance tradeoff, methods for diagnosing and preventing overfitting, and the theoretical causes and architectural solutions for vanishing gradients in deep learning.

  Machine LearningTechnical ScreenDeep LearningBias-Variance TradeoffRegularization
• PinterestEasy·Machine Learning Engineer
  Explain learning-rate fluctuation and vanishing gradients

  This question evaluates a candidate's understanding of optimization dynamics (effects of learning rate on training stability) and backpropagation-related issues (vanishing gradients) in deep neural networks, assessing competencies in training behavior and gradient propagation within the Machine Learning domain.

  Machine LearningTechnical Screen
• PinterestHard·Machine Learning Engineer
  Explain overfitting and how to prevent it

  This question evaluates understanding of overfitting, the relationship between training and validation performance and generalization, and knowledge of regularization concepts including differences between L1 and L2 penalties.

  Machine LearningTechnical Screen
• PinterestHard·Machine Learning Engineer
  Explain overfitting, underfitting, and regularization

  This question tests the conceptual understanding of the bias-variance tradeoff and the practical application of regularization to balance model complexity and generalization error.

  Machine LearningTechnical ScreenBias-VarianceRegularizationGeneralization
• PinterestHard·Machine Learning Engineer
  Implement Naive Bayes classifier from scratch

  This question evaluates understanding of probabilistic classification, specifically competence in estimating class priors, feature likelihood parameters, and computing posterior scores for Naive Bayes variants.

  Machine LearningOnline Assessment
• PinterestHard·Machine Learning Engineer
  Implement bagging with decision trees

  This question evaluates a candidate's competency in ensemble learning and practical implementation skills, including bootstrap sampling, decision tree base learners, aggregation via majority voting with tie-breaking, and managing data using pure Python lists and standard library routines.

  Machine LearningOnline Assessment
• PinterestHard·Machine Learning Engineer
  Prevent Vanishing Gradients in Deep Networks

  Tests a candidate's mathematical understanding of the vanishing gradient problem in deep neural networks and their ability to evaluate architectural mitigations including weight initialization, activation functions, residual connections, and normalization layers.

  Machine LearningTechnical ScreenDeep LearningOptimizationNeural Networks
• Point72Medium·Machine Learning Engineer
  Design Features for Residual Volatility

  This problem tests the ability to decompose financial asset returns into systematic and idiosyncratic components and design features to capture residual volatility.

  Machine LearningTechnical ScreenTime-SeriesFinanceFeature Engineering
• RedditMedium·Machine Learning Engineer
  Analyze CTR Data and Train Model

  This problem tests the end-to-end workflow of building a CTR prediction model, focusing on feature engineering for high-cardinality categorical data and handling temporal leakage.

  Machine LearningOnsiteFeature EngineeringCTR PredictionData Leakage
• RedditMedium·Machine Learning Engineer
  Build and evaluate click prediction models

  This question evaluates skills in supervised probabilistic modeling for click-through rate prediction, including model selection, calibration, evaluation metric choice, cross-validation, hyperparameter tuning, and production concerns like serving and monitoring.

  Machine LearningTechnical Screen
• RedditHard·Machine Learning Engineer
  Explain and Reduce Position Bias in Ranking

  Assess understanding of position bias in click-based ranking systems, including formalizing the examination hypothesis, estimating observation propensities, training with Inverse Propensity Scoring (IPS), and designing counterfactual evaluation metrics.

  Machine LearningTechnical ScreenRankingRecommendationPosition Bias
• RedditMedium·Machine Learning Engineer
  Model y from x and interpret distributions

  This question evaluates modeling judgment in machine learning, covering problem framing (regression versus classification), baseline model and metric selection, interpretation of feature or class distribution differences, and cold-start strategies for users, items, and regions.

  Machine LearningOnsite
• RRo.CoMedium·Machine Learning Engineer
  Debug an inaccurate patient-message triage classifier service and write up fixes

  A one-hour take-home for a senior applied AI engineering role: debug a telehealth service that sorts patient messages into care categories with an urgency and a confidence score, decide when to call an LLM or send a message to manual review, and write up bugs, priorities and trade-offs. It tests debugging, safety judgment and ML evaluation.

  Machine LearningOnline Assessment
• RRunwayHard·Machine Learning Engineer
  Implement n-gram model and select n

  This interview question evaluates core ML concepts, assumptions, math intuition, training/evaluation trade-offs, and practical failure modes in a realistic interview setting. A strong answer for Implement n-gram model and select n states assumptions, handles edge cases, explains trade-offs, and shows how to validate the result clearly.

  Machine LearningTechnical Screen
• SalesforceMedium·Machine Learning Engineer
  Explain ML and LLM fundamentals

  This question evaluates understanding of core machine learning and LLM competencies including evaluation metrics (F1 vs accuracy), differences between classification and regression, fair benchmarking, inference pipeline bottlenecks and optimization, productionization of AI services, and modern LLM concepts such as the Transformer architecture, context engineering, retrieval-augmented generation, grounding, and guardrails. It is commonly asked in Machine Learning interviews to assess both conceptual understanding and practical application, probing the candidate's ability to reason about evaluation trade-offs, system-level performance constraints, and deployment considerations rather than only algorithmic detail.

  Machine LearningOnsite
• SamsaraMedium·Machine Learning Engineer
  Explain LLM training and evaluation

  This question evaluates an engineer's competency in LLM engineering, including transformer pretraining and finetuning, alignment approaches like instruction tuning and RLHF, strategies for reducing hallucinations, evaluation and production monitoring practices, and inference optimizations.

  Machine LearningOnsite
• Scale AIEasy·Machine Learning Engineer
  Explain LLM post-training methods and tradeoffs

  This question assesses the candidate's understanding of the LLM post-training pipeline, specifically how to transition a foundation model into a helpful, safe assistant through various alignment techniques.

  Machine LearningOnsiteLLMAlignmentRLHF
• Scale AIHard·Machine Learning Engineer
  Explain Transformers, attention, decoding, RL, and evaluation

  This question evaluates understanding of transformer architectures, self- and cross-attention mechanisms, autoregressive decoding and sampling strategies, reinforcement learning–based fine-tuning (RLHF), evaluation methodologies, and training/inference optimization techniques.

  Machine LearningTechnical Screen
• Scale AIMedium·Machine Learning Engineer
  Implement universal adversarial attack on GPT-2

  Machine learning robustness prompt on safely evaluating universal adversarial prompts for GPT-2, covering threat modeling, differentiable optimization concepts, transferability, metrics, safety controls, and mitigation strategies without deployment-ready jailbreak details.

  Machine LearningTechnical Screen
• SSealthEasy·Machine Learning Engineer
  Represent k-means as an MLP

  This question tests the conceptual mapping between the distance-based assignment step of k-means clustering and a single-layer neural network architecture.

  Machine LearningOnsiteK-MeansNeural NetworksLinear Algebra
• ShopifyMedium·Machine Learning Engineer
  Build a fraud detection model

  Design an end-to-end machine learning system for real-time transaction fraud detection, covering label formulation under delayed feedback, feature design, threshold selection, and continuous monitoring.

  Machine LearningTechnical Screenfraud-detectionmachine-learningsystem-design
• ShopifyMedium·Machine Learning Engineer
  Build model to predict package delivery time

  This question evaluates a candidate's competence in end-to-end machine learning system design for time-to-event forecasting, covering prediction target definition and timing, feature engineering and model selection, handling missing or delayed signals and data leakage, uncertainty quantification, and production monitoring and retraining.

  Machine LearningOnsite
• ShopifyMedium·Machine Learning Engineer
  Model Product Ranking

  This question evaluates understanding of end-to-end machine learning ranking systems, including label definition, feature engineering, model selection, training strategies, and offline/online evaluation.

  Machine LearningOnsite