Viva prep · Real questions · Student experiences Enroll in Bootcamp

Proctor level1_viva4

All approved submissions for this examiner code.

level1_viva4 MLP Level 1 Aug 5, 2026
  1. 1
    Tell me about yourself.
  2. 2
    What is your Kaggle leaderboard score/rank?
  3. 3
    Explain your notebook/project.
  4. 4
    Explain your EDA.
  5. 5
    How many graphs did you create?
  6. 6
    Explain each graph you created.
  7. 7
    What observations did you make from your graphs?
  8. 8
    How do you hide/show values in a correlation heatmap?
  9. 9
    Explain your boxplot.
  10. 10
    Explain your density regression plot.
  11. 11
    What feature engineering did you perform?
  12. 12
    Why did you perform that feature engineering?
  13. 13
    How did you handle missing values?
  14. 14
    How did you handle numerical missing values?
  15. 15
    How did you handle categorical missing values?
  16. 16
    How did you handle disguised missing values (None, NaN, Unknown, Unspecified)?
  17. 17
    Why did you preserve the "Unknown" category?
  18. 18
    What numerical transformers did you use?
  19. 19
    What categorical transformers did you use?
  20. 20
    Why did you use those transformers?
  21. 21
    What is Data Leakage?
  22. 22
    When does Data Leakage occur?
  23. 23
    How did you prevent Data Leakage?
  24. 24
    What is an N-gram?
  25. 25
    What is Unigram?
  26. 26
    What is Bigram?
  27. 27
    What is a 6-gram, 8-gram, or 11-gram?
  28. 28
    What is TF-IDF?
  29. 29
    Difference between word-level and character-level TF-IDF.
  30. 30
    What does TF-IDF Vectorizer do?
  31. 31
    Explain the parameters of TF-IDF Vectorizer.
  32. 32
    Why did you choose 50,000 TF-IDF features?
  33. 33
    What models did you use?
  34. 34
    Why did you choose those models?
  35. 35
    What is your best model?
  36. 36
    Why is it your best model?
  37. 37
    What is an Ensemble model?
  38. 38
    Explain your final modeling approach.
  39. 39
    What hyperparameter tuning did you perform?
  40. 40
    Which hyperparameters did you tune?
  41. 41
    How many model fits will GridSearchCV perform?
  42. 42
    How many model fits will RandomizedSearchCV perform?
  43. 43
    Difference between GridSearchCV and RandomizedSearchCV.
  44. 44
    How do you know how many fits will be performed?
  45. 45
    What is n_estimators?
  46. 46
    Why did you choose n_estimators = 100?
  47. 47
    What happens if n_estimators = 2000?
  48. 48
    What is max_depth?
  49. 49
    What is learning_rate?
  50. 50
    What is n_iter?
  51. 51
    What would you do if the model overfits because of max_depth?
  52. 52
    How do you prevent overfitting?
  53. 53
    How do you improve an underfitting model?
  54. 54
    What is IQR?
  55. 55
    How does outlier detection help?
  56. 56
    How does a boxplot help detect outliers?
  57. 57
    How does ColumnTransformer work?
  58. 58
    Why didn't you use scaling?
  59. 59
    Should SGD be used without scaling?
  60. 60
    What is SGD?
  61. 61
    Why did you use SGD?
  62. 62
    Difference between Accuracy, Precision, Recall, and F1-Score.
  63. 63
    Why did you use F1-Score?
  64. 64
    What is RMSLE?
  65. 65
    What is Blended Score?
  66. 66
    Why did your RMSLE score decrease after blending?
  67. 67
    What is a Decision Tree?
  68. 68
    What algorithm do XGBoost, LightGBM, and CatBoost use?
  69. 69
    Difference between XGBoost, LightGBM, and CatBoost.
  70. 70
    Are these models linear or non-linear?
  71. 71
    Explain your complete preprocessing and modeling pipeline.
  72. 72
    How does your complete pipeline work?
  73. 73
    What are custom classes?
  74. 74
    Why did you create custom classes?
  75. 75
    Did you use any LLMs while building the project?
  76. 76
    Coding: Impute the most frequent categorical value for each column.
  77. 77
    Coding: Perform PCA on numerical features and obtain the top two components.
  78. 78
    Coding: Perform K-Means clustering on a dataset.
  79. 79
    Coding: Load a dataset, preprocess it, perform train-test split, and build a Linear Regression model.
  80. 80
    Coding: Implement RandomizedSearchCV.
  81. 81
    Coding: Create a complete Pipeline with ColumnTransformer, Imputers, Transformers, and a model.
level1_viva4 MAD1 Level 1 Jul 30, 2026
  1. 1
    Perform the checksum verification.
  2. 2
    Demonstrate the core functionalities of the application.
  3. 3
    Show and explain your models.
  4. 4
    Which database did you use in your project?
Created for educational purposes only. Questions are based on students' personal experiences and may not reflect actual exam content.