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

What the job involves

As a Data Scientist at an AI and deeptech company like Arya.ai, you will build and deploy machine learning models for domains such as fintech, insurance, and enterprise automation. Day to day involves data cleaning, feature engineering, model training and evaluation, and collaborating with engineers to productionise models via APIs. You will also monitor model performance and iterate based on business metrics.

Role and Technical Knowledge

1. 1

   Explain the bias-variance tradeoff and how you would address high bias or high variance in a model.

   Define bias and variance, relate to underfitting and overfitting, and mention techniques like regularization, cross-validation, or ensemble methods.

2. 2

   How would you handle missing data in a dataset for a credit risk model?

   Discuss deletion, imputation (mean, median, mode, or model-based), and the importance of understanding missingness patterns and business context.

3. 3

   Compare and contrast Random Forest and Gradient Boosting. When would you use each?

   Explain bagging vs boosting, interpretability, training speed, and performance. Mention scenarios like small data vs large data, and need for explainability.

4. 4

   What is regularization? Explain L1 and L2 and their use cases.

   Describe Lasso and Ridge, how they penalize coefficients, and when to use L1 for feature selection or L2 for multicollinearity.

Practical Scenarios and Case Studies

1. 5

   You are given a dataset of customer transactions and asked to build a fraud detection model. How would you approach it?

   Cover data exploration, class imbalance handling, feature engineering (time, frequency, amount), model choice (e.g., anomaly detection, tree-based), and evaluation metrics like precision-recall AUC.

2. 6

   How would you deploy a machine learning model as a REST API for a real-time prediction service?

   Mention frameworks like Flask or FastAPI, containerization with Docker, and considerations for latency, scaling, and monitoring.

3. 7

   Your model's performance drops after deployment. How would you debug and address it?

   Discuss data drift, concept drift, logging, retraining, and setting up monitoring for input and output distributions.

4. 8

   Explain how you would design an A/B test to measure the impact of a new recommendation algorithm.

   Cover hypothesis formulation, randomization, sample size calculation, metric selection, and statistical significance testing.

Behavioural and HR Questions

1. 9

   Why do you want to work at a fast-growing AI deeptech startup like Arya.ai?

   Show passion for applied AI, mention the company's domain focus, and align with your desire for rapid learning and impact.

2. 10

   Describe a time when you had to learn a new technology quickly for a project.

   Use STAR method, highlight self-learning, resourcefulness, and the outcome.

3. 11

   Tell me about a time you worked in a team to solve a complex problem.

   Focus on collaboration, communication, and your specific contribution. Mention any conflict resolution.

4. 12

   Where do you see yourself in 3-5 years in the field of data science?

   Show ambition to grow into a lead or specialist role, with continued learning in AI and deeptech.

Coding and Problem Solving

1. 13

   Write a Python function to calculate the F1 score given true and predicted labels.

   Implement precision, recall, and harmonic mean. Handle division by zero and edge cases.

2. 14

   Given a list of numbers, find the median without using built-in functions.

   Sort the list, handle even and odd lengths, and consider time complexity.

3. 15

   Write a SQL query to find the second highest salary from an Employee table.

   Use subquery with LIMIT/OFFSET or window functions like DENSE_RANK. Explain the logic.

4. 16

   Implement a simple linear regression using gradient descent in Python.

   Define hypothesis, cost function, and update rule. Show iterative optimization and convergence check.