Zomato | Data Scientist | Dec'22 [Offer]

zomato logo
zomato
data scientist2 yearsOffer
January 9, 20230 reads

Summary

I was offered a Data Scientist role at Zomato after undergoing multiple rounds, including managerial, technical, and case study assessments. The process spanned several weeks, culminating in a compensation discussion and a final offer.

Full Experience

0. Application

I set my LinkedIn status to "open to work," and within a week, I received an inmail from Zomato's VP of Data Science, expressing interest in my profile for a Data Scientist role. I later learned that their profile is managed by recruiters. I responded with a **yes**.

1. First Round - Managerial (20 mins)

This round was with the VP, who discussed the company culture, ongoing projects, and their expectations for the role. I was asked about my approach to solving data science problems, which led to cross-questioning. We also discussed my past work experience in my current and previous companies, followed by further cross-questioning.

2. Second Round - Technical (1 Hr)

A senior data scientist conducted this round. I was asked questions related to my CV, such as explaining a project end-to-end. The interviewer repeatedly questioned my choice of approaches in projects. I also had to explain the working of various machine learning algorithms and methods, including how to train a Word2Vec model and the concepts of bagging and boosting. There were a few basic statistics questions as well. A significant case study involved discussing how I would build an ETA calculation model for Zomato. No questions were asked about deep learning. I managed to answer most questions but made mistakes on 2-3 of them. For the case study, it was an open-ended discussion where they focused on my problem-solving approach.

3. Third Round - Technical and Case Study (1 Hr)

This round was with a principal data scientist. Interestingly, some questions I had struggled with in the previous round were asked again. This time, I provided the correct answers, and the interviewer seemed impressed by my ability to learn from my mistakes and improve before the next stage. We again discussed common machine learning concepts. I was then presented with three data science case study problems, where the goal was to assess my ability to structure problems and solutions, identify relevant variables, plan data collection, and determine retraining frequency. Specifically, I was asked:

  • How to design a recommender system for Zomato to suggest dishes to a user.
  • How I would solve the cold-start problem in such a system.
  • What metrics I would choose to evaluate the performance of my solution.

4. Compensation Discussion Round

The VP conducted this final round. They provided feedback from the technical rounds, acknowledging my improvement from the second to the third round. While they found my knowledge and problem-solving skills strong, they noted that my prior projects might not be directly applicable, implying a significant learning curve upon joining. Subsequently, they extended an offer: 24LPA Fixed and 8Lacs worth of ESOPS vested annually.

Interview Questions (6)

Q1
How to Train Word2Vec
Data Structures & Algorithms

Explain the process of training a Word2Vec model.

Q2
Explain Bagging and Boosting
Data Structures & Algorithms

Explain the concepts of Bagging and Boosting in machine learning.

Q3
Design ETA Calculation Model for Zomato
System Design

Describe your approach to building an Estimated Time of Arrival (ETA) calculation model specifically for Zomato, considering real-world factors.

Q4
Design Zomato Dish Recommender System
System Design

Design a recommender system for Zomato to suggest dishes to users. Discuss how to structure the problem, identify variables, collect data, and determine retraining frequency.

Q5
Solve Cold-Start Problem in Recommender Systems
System Design

Explain how you would address the cold-start problem in the context of a recommender system.

Q6
Metrics for Recommender System Performance
System Design

What metrics would you choose to evaluate the performance of a recommender system?

Preparation Tips

My preparation involved several online courses, with the following being my key recommendations:

  1. CS229 from YouTube: Taught by either Andrew NG or Anand Avati.
  2. Python Bootcamp - Udemy: Taught by Jose Portilla.
  3. Data Science and Machine learning - Udemy: Taught by Jose Portilla.
  4. Fast AI machine learning course: Available on YouTube.
  5. Deeplearning.ai specialization: Although no deep learning questions were asked, it's still valuable knowledge.
  6. Statquest by Josh Starmer: I consider this the best lecture series available for statistics.
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!