Niki.ai Data Scientist Interview: Rounds, Questions and Preparation

What the job involves

As a Data Scientist at an AI-first startup like Niki.ai, you will build and deploy machine learning models that power conversational AI, recommendation systems, and predictive analytics. Day to day, you will work with messy real-world data, run experiments, and collaborate with engineers to ship models into production. Expect a mix of data cleaning, model training, evaluation, and rapid iteration in a fast-paced environment.

Role and technical knowledge

1. 1

   Explain the difference between supervised and unsupervised learning. Give an example relevant to a conversational AI product.

   Cover definitions, key algorithms, and use cases like intent classification (supervised) vs. topic clustering from user queries (unsupervised).

2. 2

   How would you handle missing values in a dataset for a recommendation model?

   Discuss imputation methods (mean, median, model-based), dropping rows, and the impact on model bias. Mention that the choice depends on missingness pattern.

3. 3

   What is overfitting and how do you prevent it?

   Explain overfitting as low bias high variance. Mention regularization, cross-validation, early stopping, and more data.

4. 4

   Describe how a confusion matrix works and when you would use precision over recall.

   Define TP, FP, FN, TN. Give examples: precision when false positives are costly (spam detection), recall when false negatives are costly (disease detection).

Practical scenarios and problem solving

1. 5

   You are given a dataset of user chats with a chatbot. How would you build a model to detect user intent?

   Outline steps: text preprocessing, tokenization, feature extraction (TF-IDF or embeddings), choosing a classifier (logistic regression, SVM, or BERT), training, and evaluation with F1-score.

2. 6

   How would you evaluate the performance of a recommendation system in production?

   Mention offline metrics (precision@k, recall@k, NDCG) and online metrics (click-through rate, conversion rate, user engagement). Also discuss A/B testing.

3. 7

   Suppose your model's accuracy drops after deployment. What steps would you take?

   Check data drift, concept drift, pipeline errors, and retrain with recent data. Monitor input distributions and set up alerts.

4. 8

   You need to explain a complex model's prediction to a non-technical stakeholder. How would you do it?

   Use SHAP or LIME for local explanations, or simple surrogate models. Focus on actionable insights, not technical jargon.

Behavioural and situational questions

1. 9

   Tell me about a time you worked on a team project where you had to explain a technical concept to a non-technical teammate.

   Use STAR method. Highlight empathy, clarity, and how you adapted your communication.

2. 10

   Describe a situation where you had to learn a new tool or technique quickly for a project.

   Show self-learning ability, resourcefulness, and application of the new skill to deliver results.

3. 11

   How do you handle feedback on your work, especially if you disagree?

   Demonstrate openness, ask clarifying questions, and focus on data-driven discussion rather than ego.

4. 12

   Give an example of a time you failed at something. What did you learn?

   Choose a genuine failure, take responsibility, and emphasise the lesson and subsequent improvement.

HR and motivation questions

1. 13

   Why do you want to work at a fast-growing AI startup like Niki.ai rather than a large company?

   Show enthusiasm for ownership, rapid iteration, and direct impact. Mention specific aspects of the startup's work if known.

2. 14

   Where do you see yourself in three years?

   Align your growth with the company's needs: deepening ML expertise, leading projects, or mentoring. Avoid saying you want to leave for a bigger company.

3. 15

   What are your salary expectations?

   Research typical fresher salaries in Indian AI startups. Give a range and express flexibility based on overall compensation and learning opportunities.

4. 16

   Do you have any questions for us?

   Ask about team structure, tech stack, mentorship, or upcoming projects. Shows genuine interest and helps you evaluate fit.