Swiggy | Data Scientist 1 | Interview

swiggy logo
swiggy
Data Scientist 1Bangalore
May 13, 20257 reads

Summary

I successfully cleared an off-campus interview process for a Data Scientist 1 role at Swiggy, comprising four distinct rounds: a coding assessment with DSA and Python fundamentals, an in-depth ML and statistics discussion, a challenging ML system design round, and a final behavioral hiring manager interview, ultimately leading to an offer.

Full Experience

My Interview Experience and offer (Swiggy Data Scientist 1 role)

Off-campus opportunity via referral. All of these were knockout rounds

1) Coding round:

  • 1 hr interview with data scientist 2
  • The coding-focused interview covering Python (functions, try-except, loops, decorators, multithreading), basic Pandas (like groupby), and SQL (including subqueries). It also included one DSA question—Stocks Buy and Sell (LeetCode, easy-medium level). Throughout the round, I was asked to explain my approach and reasoning behind each solution.

2) ML depth/breath round:

  • 1 hr interview with senior data scientist
  • The ML depth/breath round covered core statistics including mean, variance, central limit theorem, Type I/II errors, the reasoning behind sample standard deviation being σ/√n, and a proof for the variance of the sample mean. Probability questions were also asked. In data science, I was questioned on evaluation metrics like precision, recall, F1-score, ROC-AUC, regularization techniques, bias-variance tradeoff, and metrics for evaluating KMeans clustering. The deep learning section included CNN output size formula with reasoning behind each term (like why +1 is added), dropout behavior (including impact of stacking dropout layers vs placing dropout between layers), vanishing gradients in RNNs, and how LSTMs solve them.

P.S. This is just summary of second round. He dived deep into each concepts and created "what if" scenarios

3) Problem Solving round:

  • 1 hr interview with Director Instamart (they were hiring for Instamart team)
  • The Problem Solving round focused on ML system design, starting with a discussion on my past work, approach, and reasoning. I was then given real-world problems the team is working on. The first task was to design a solution for dynamically displaying filters based on the searched product. The second was to build a model to estimate delivery time, inspired by a Swiggy Bytes technical blog.

4) Hiring Manager round:

  • 1 hr interview with Senior Manager Instamart
  • It was a casual conversation focused on behavioral and situational questions, discussing how I relate to Swiggy's values. It was mostly informal with no technical questions asked.

Compensation details : https://leetcode.com/discuss/post/6740557/swiggy-data-scientist-1-bangalore-by-kus-fat9/

Interview Questions (7)

Q1
Stocks Buy and Sell
Data Structures & AlgorithmsMedium

Implement the 'Stocks Buy and Sell' problem, typically found on LeetCode. I was asked to explain my approach and reasoning behind the solution.

Q2
Core Statistics and Probability Concepts
Other

Explain concepts such as mean, variance, central limit theorem, Type I/II errors, the reasoning behind sample standard deviation being σ/√n, and provide a proof for the variance of the sample mean. Also, answer general probability questions.

Q3
ML Evaluation Metrics and Techniques
Other

Discuss evaluation metrics such as precision, recall, F1-score, ROC-AUC. Explain regularization techniques and the bias-variance tradeoff. Describe metrics used for evaluating KMeans clustering.

Q4
Deep Learning Concepts
Other

Explain the CNN output size formula with reasoning for each term. Discuss dropout behavior, including the impact of stacking dropout layers versus placing dropout between layers. Describe vanishing gradients in RNNs and how LSTMs solve them.

Q5
ML System Design: Dynamic Filter Display
System DesignHard

Design a solution for dynamically displaying filters based on the searched product. This was framed as a real-world problem the team is working on, and I started with a discussion on my past work, approach, and reasoning.

Q6
ML System Design: Delivery Time Estimation Model
System DesignHard

Build a model to estimate delivery time, inspired by a Swiggy Bytes technical blog. This was framed as a real-world problem the team is working on, and I started with a discussion on my past work, approach, and reasoning.

Q7
Behavioral and Situational Questions
Behavioral

Answer behavioral and situational questions focused on how I relate to Swiggy's values. This was mostly an informal conversation with no technical questions asked.

Discussion (0)

Share your thoughts and ask questions

Join the Discussion

Sign in with Google to share your thoughts and ask questions

No comments yet

Be the first to share your thoughts and start the discussion!