Curated interview practice questions with answers. These are practice scenarios, not a verified list of past interview questions.
About Paytm hiring for Machine Learning roles
Paytm, operated by One 97 Communications, is an Indian digital payments and financial services company offering UPI payments, merchant soundboxes, payment gateway, lending distribution and ticketing. Its technology teams work on payments infrastructure, risk and merchant services. It hires for software engineers, QA engineers, data analysts and data scientists.
For Machine Learning roles, this Paytm practice set focuses most on Decision Trees and Ensembles (7), Classification Models (5), Model Evaluation and Tuning (4), Clustering and Dimensionality Reduction (4). It has 1 easy, 21 medium, 8 hard questions, so you can start easy and work up to the harder ones.
Company type
Fintech
Headquarters
Noida, India
Main India offices
Noida, Bengaluru, Mumbai
Paytm Machine Learning interview process
Online assessment (coding + MCQs)
Technical round 1
Technical round 2
Hiring manager round
HR round
Typical pattern based on public candidate experiences. Rounds vary by role, team and year.
Solved for PaytmStart practicing
3 QUESTIONS FREESign up to unlock the full question bank and track your progress.Unlock every question and answer, then continue practising company-wise and package-wise.
✓ 3 questions free • Unlock the rest in seconds
01
Check calibration with a reliability curve and Brier score.
from sklearn.calibration import CalibrationDisplay
from sklearn.metrics import brier_score_loss
proba = clf.predict_proba(X_test)[:, 1]
print('Brier', brier_score_loss(y_test, proba))
CalibrationDisplay.from_predictions(y_test, proba, n_bins=10)
plt.show()
Points on the diagonal mean predicted probabilities match observed frequencies.
02
How do you choose a classification threshold that maximises F1?
from sklearn.metrics import precision_recall_curve
import numpy as np
proba = clf.predict_proba(X_val)[:, 1]
p, r, t = precision_recall_curve(y_val, proba)
f1 = 2 * p * r / (p + r + 1e-9)
best = t[np.argmax(f1[:-1])]
print(best)
Choose it on validation data, not the test set. scikit-learn 1.5+ also offers TunedThresholdClassifierCV.
import lightgbm as lgb
m = lgb.LGBMClassifier(n_estimators=5000, learning_rate=0.05,
num_leaves=31, subsample=0.8, subsample_freq=1, colsample_bytree=0.8)
m.fit(X_train, y_train, eval_set=[(X_val, y_val)],
eval_metric='auc', callbacks=[lgb.early_stopping(100)])
print(m.best_iteration_)
The large n_estimators is only an upper limit; early stopping picks the real number.
04
How do one-vs-rest and multinomial approaches differ for multiclass classification?
One-vs-rest trains one binary classifier per class against all others and picks the highest score; it is simple but probabilities may not sum to one. Multinomial (softmax) logistic regression models all classes jointly with a single loss, giving consistent probabilities. For metrics, use macro averaging to treat classes equally or weighted averaging to account for class frequency.
05
A demand forecast is consistently under-predicting during festivals. How would you fix it?
The model lacks festival information. I would add features for festival dates such as Diwali or regional holidays, days-to-festival, and past festival uplift for the same item. If the business cost of stock-outs is higher than over-stock, I would use a quantile loss for a higher percentile. Then validate specifically on past festival periods, not just the average error.
06
What fairness checks would you run on a model that affects people?
Compare metrics across groups such as gender, age band or region: approval rate (demographic parity), true positive rate (equal opportunity) and false positive rate. Check calibration per group too. Remove or audit proxy features like pin code that may stand in for protected attributes. Document findings in a model card, and involve legal or compliance teams, especially for lending and hiring use cases.
from sklearn.model_selection import LearningCurveDisplay
import numpy as np
LearningCurveDisplay.from_estimator(model, X, y, cv=5,
train_sizes=np.linspace(0.1, 1.0, 6), scoring='roc_auc', n_jobs=-1)
plt.show()
A persistent gap between curves suggests more data or regularisation; two low curves suggest a richer model or better features.
08
Your gradient boosting model performs worse on recent months than older ones. What do you check?
I would check for drift between periods in feature distributions and target rate, and confirm that the evaluation uses a time-based split. If behaviour has changed, I would train on a more recent window or weight recent data more, add features that capture the change, and set up a regular retraining schedule. Comparing SHAP importance across periods shows which relationships have shifted.
09
How do you evaluate clustering when there are no labels?
Use internal metrics: silhouette score, Davies-Bouldin index (lower is better) and Calinski-Harabasz index (higher is better). Check stability by re-clustering on bootstrap samples. Most important is external usefulness: profile each cluster with key statistics and check whether business teams find the segments distinct and actionable, for example different response to a campaign.
10
Explain a single prediction with SHAP for a tree model.
import shap
explainer = shap.TreeExplainer(model)
sv = explainer(X_val)
shap.plots.waterfall(sv[0])
shap.plots.beeswarm(sv)
The waterfall shows how each feature pushes one prediction up or down from the base value; the beeswarm shows global importance and direction. TreeExplainer is fast and exact for tree ensembles.
11
Your dataset has a city column with 2,000 unique values. How will you encode it?
One-hot would create 2,000 sparse columns, so I would first group rare cities below a frequency threshold into 'Other'. Then options are frequency encoding, out-of-fold target encoding, or mapping cities to useful attributes such as tier, state or population. For boosting models like CatBoost or LightGBM, native categorical handling also works well. I would compare options with cross-validation.
12
Compare filter, wrapper and embedded feature selection methods.
Filter methods score features independently of a model using statistics such as correlation, chi-square or mutual information; they are fast but ignore interactions. Wrapper methods like Recursive Feature Elimination train models on feature subsets; they are accurate but slow. Embedded methods select during training, such as L1 regularisation or tree feature importances. Embedded methods are a common practical middle ground.
13
Your manager asks for a single number showing the model's business value. What do you give?
I would convert model outcomes into money or time using a cost matrix agreed with the business, for example savings from caught fraud minus the cost of reviewing false alerts. I would compare this against the current process or a rule-based baseline on holdout data and show a range rather than one point. A short note on assumptions keeps the number honest.
Isolation Forest builds random trees that split on random features and random thresholds. Anomalies are few and different, so they get isolated in fewer splits, giving shorter average path lengths. The anomaly score comes from that path length. It is fast, works without labels and scales well. The contamination parameter sets the expected share of anomalies and determines the threshold.
15
A new model version performs badly right after release. How do you handle it?
Roll back immediately to the previous version from the model registry, which should be a one-step action. Then investigate: compare the input features served with those used in validation, check for training-serving skew, version mismatches or a wrong artifact. Add the failure case to automated pre-release checks, and use canary or shadow releases next time so issues affect only a small share of traffic.
16
What is association rule mining? Explain support, confidence and lift.
Association rule mining finds items that occur together, such as in market basket analysis. Support is how often an itemset appears in all transactions. Confidence of A to B is the share of transactions with A that also contain B. Lift is confidence divided by B's support; lift above 1 means A and B occur together more than by chance. Apriori and FP-Growth are common algorithms.
17
You must build a spam classifier quickly with limited compute. What do you choose?
I would use TF-IDF features with a Multinomial Naive Bayes or linear model like logistic regression or LinearSVC. They train in seconds, handle high-dimensional sparse text well and give strong baselines. I would evaluate precision carefully because blocking genuine mail is costly. Only if this is not good enough would I try fine-tuning a small transformer model.
18
Random forest and LightGBM give the same AUC. Which would you deploy?
I would compare beyond AUC: model size, inference latency, training time, calibration and stability across time periods. LightGBM is usually smaller and faster to score, while random forest needs less tuning and is less prone to overfitting. If both are equal, I would choose the one that is cheaper to run and easier for the team to maintain and retrain.
The kernel trick computes dot products of data in a high-dimensional feature space without explicitly transforming the data, using a kernel function such as RBF, exp(-gamma times squared distance). This lets linear algorithms like SVM learn non-linear boundaries efficiently. In RBF SVMs, larger gamma makes each point's influence more local, which can lead to overfitting.
20
Tell me about a time you had to explain a model's error to a business team.
Situation: a sales forecast missed by 18 percent in one month. Task: explain why to the category team. Action: I broke the error down by region and found most came from a sudden promotion not in our data. I showed a simple chart and proposed adding a promotion calendar feed. Result: the team shared promotion plans in advance, and the next quarter's error fell below 8 percent.
from sklearn.cluster import DBSCAN
import numpy as np
db = DBSCAN(eps=0.5, min_samples=10).fit(Xs)
labels = db.labels_
n_clusters = len(set(labels)) - (1 if -1 in labels else 0)
print(n_clusters, 'clusters,', np.sum(labels == -1), 'noise')
Pick eps from the knee of a sorted k-nearest-neighbour distance plot.
22
What is the difference between a generative and a discriminative model?
A discriminative model learns the decision boundary or P(y|x) directly, such as logistic regression or SVM. A generative model learns how data is produced, P(x|y) and P(y), and can generate new samples, such as Naive Bayes or Gaussian mixture models. Discriminative models usually give better accuracy with enough data, while generative models handle missing data and small samples better.
23
A model's accuracy dropped 10 percent over three months in production. How do you investigate?
I would first check data pipelines for broken features, nulls or schema changes, then compare current feature distributions with training data using PSI to find drift. I would check whether the label definition or business process changed, and analyse errors by segment. If drift is real, retrain on recent data, possibly with time-weighted samples, and set up automatic drift alerts and scheduled retraining.
24
When would you use AdaBoost and how does it re-weight samples?
AdaBoost trains weak learners, usually decision stumps, in sequence. After each round it increases the weights of misclassified samples so the next learner focuses on them, and gives each learner a vote weight based on its accuracy. It works well on clean data but is sensitive to noisy labels and outliers, which get ever larger weights. Gradient boosting is a generalisation that has largely replaced it.
25
Use HistGradientBoosting with native categorical support in scikit-learn.
from sklearn.ensemble import HistGradientBoostingClassifier
clf = HistGradientBoostingClassifier(
categorical_features='from_dtype', max_iter=500,
learning_rate=0.05, early_stopping=True, random_state=42)
X_train[cat_cols] = X_train[cat_cols].astype('category')
clf.fit(X_train, y_train)
It handles missing values natively and is much faster than GradientBoostingClassifier on large data.
Bias is error from overly simple assumptions, which causes underfitting. Variance is error from sensitivity to the particular training sample, which causes overfitting. Increasing model complexity usually lowers bias but raises variance. The goal is the complexity that minimises total test error. For example, a linear model on curved data has high bias, while a very deep unpruned tree has high variance.
27
The fraud team can only review 500 alerts a day. How do you set the threshold?
Instead of 0.5, rank transactions by predicted probability and pick the threshold that yields about 500 alerts per day on recent data. Then report precision and recall at that cut-off, or precision at k. I would monitor daily alert volume because drift can change it, and recompute the threshold periodically. If cost data exists, I would choose the threshold that maximises expected savings.
28
Use statsmodels to get p-values for regression coefficients.
import statsmodels.api as sm
Xc = sm.add_constant(X_train)
ols = sm.OLS(y_train, Xc).fit()
print(ols.summary())
The summary shows coefficients, standard errors, p-values, confidence intervals, R-squared and diagnostics. scikit-learn focuses on prediction and does not report p-values, so statsmodels is preferred for inference.
29
Which hyperparameters matter most when tuning a gradient boosting model?
The key ones are learning_rate and number of trees, tuned together with early stopping; tree complexity through max_depth or num_leaves and min_child_samples; and randomness through subsample and colsample_bytree. Regularisation terms like lambda and alpha help on noisy data. A common approach is to fix a small learning rate such as 0.05, use early stopping for trees, then tune complexity.
30
Why do skewed features often need a log or power transform?
Heavily right-skewed features like income or transaction amount have long tails that can dominate linear models and distance metrics. A log1p, Box-Cox or Yeo-Johnson transform compresses large values and makes the distribution closer to normal, which often improves linear and distance-based models. Yeo-Johnson works with zero and negative values. Tree models usually do not need these transforms.
Tips to crack the Paytm Machine Learning interview
Learn UPI flow basics, including the role of NPCI, PSP apps and banks, for payments domain questions.
Practise testing scenarios for payment failures, refunds and timeouts if applying for QA roles.
Read about Paytm's merchant business like Soundbox and payment gateway, which are major parts of its strategy.
Keep a one-minute story ready for every project on your CV: problem, your part, tools, result.
Practise the hard questions on this page aloud; Paytm interviewers usually ask follow-ups on your first answer.
Frequently asked questions
How many rounds are there in the Paytm Machine Learning interview?
Candidates usually go through 5 stages: Online assessment (coding + MCQs); Technical round 1; Technical round 2; Hiring manager round; HR round. The exact rounds change with the role, team, location and your experience, so confirm the process with the recruiter.
Which topics should I prepare for the Paytm Machine Learning interview?
This practice set leans on Decision Trees and Ensembles (7), Classification Models (5), Model Evaluation and Tuning (4), Clustering and Dimensionality Reduction (4). Revise those first, then practise explaining one or two of your own projects end to end, because most Paytm interviewers ask you to walk through real work.
Are these the exact questions asked at Paytm?
No. These are curated practice questions that match the skills and difficulty usually tested in fintech interviews like Paytm's. They are not a leaked or verified list of past questions.
Where does Paytm hire Machine Learning professionals in India?
Paytm has large teams in Noida, Bengaluru, Mumbai. Openings and locations change often, so check the Paytm careers page and filter by city and experience.