Data & Machine Learning

ML Fundamentals

18 practice questions. Free questions open a full answer guide; the rest unlock with Pro.

  • What is overfitting, and what are three techniques to prevent it? Junior level
  • How does cross-validation work, and what are the tradeoffs between k-fold and leave-one-out cross-validation?Go Pro Mid level
  • What is the difference between L1 and L2 regularisation? When would you prefer each?Go Pro Mid level
  • Explain how you would handle a highly imbalanced classification dataset (e.g., 1% positive class).Go Pro Mid level
  • How does gradient boosting differ from random forests, and when would you choose one over the other?Go Pro Mid level
  • A model you shipped six months ago is gradually losing accuracy in production. How do you investigate what's going wrong and decide what to do about it?Go Pro Mid level
  • What is data leakage in machine learning, and how would you detect and prevent it?Go Pro Mid level
  • What is the difference between supervised and unsupervised learning? Give one real-world use case for each.Go Pro Junior level
  • Your training data has 98% non-fraud and 2% fraud cases. How would you approach training a classifier on this imbalanced dataset?Go Pro Junior level
  • What is the bias-variance tradeoff, and how does it affect model selection?Go Pro Junior level
  • Your organisation wants to adopt responsible AI practices. How do you operationalise fairness and bias testing for ML models?Go Pro Senior level
  • A junior data scientist comes to you with a model that has 98% accuracy but you suspect it's flawed. How do you mentor them through diagnosing it?Go Pro Senior level
  • How do you manage technical debt in a ML codebase that has grown organically over three years with multiple contributors?Go Pro Senior level
  • Walk me through how you version and reproduce a model training run for auditability.Go Pro Mid level
  • A model that performed well at launch is now quietly degrading in production. How do you detect this and decide when to retrain?Go Pro Senior level
  • You're building a classifier for a rare event that occurs in about 1% of cases. How do you approach the class imbalance?Go Pro Mid level
  • You are asked to reduce model inference latency by 60% without retraining. What techniques would you apply?Go Pro Senior level
  • Your model performs well in staging but degrades significantly in production within two weeks. How do you diagnose it?Go Pro Mid level
Want questions matched to your role? Paste a job title, job description, or CV for a personalized set, or go Pro to unlock the full bank.