Summary
I successfully interviewed for an SDE 2 (Data Engineering) role at Indeed, navigating through multiple rounds including DSA, SQL, System Design, Data Modelling, and Data Quality. I received an offer for the position.
Full Experience
Special Appreciation to the panels and the recruiters as they were by far the best people I have faced in my interview process journey.
Phone Screening 1 (KARAT): 45 mins
A DSA question based on hashmap and strings. Basically you were given recipes and raw materials and you need to find the dishes that can be prepared. Leetcode medium level.
2 sql questions which required knowledge of partioning and dense ranking.
Phone Screening 2 (KARAT): 45 mins
2 DSA questions were asked in this round. One was based on BFS similar to rotten oranges and 2nd question was based on the course schedule 2 with a minor tweak.
2 sql questions which required knowledge of partioning and dense ranking.
Onsite Rounds
Resume Deep Dive: 1 hour
This was taken by a hiring manager in which questions related to my resume and in depth spark applications were asked. Questions were mostly related to my proud projects and how i implemented the project along with few improvements and deep analysis on the various databases and use cases.
High Level Design 1 hour
This round was taken by a principal architect and a staff engineer where we discussed the creation of pipeline if one source is a database table and other source is API. I discussed the lambda architecture where we can have slow and fast path. The interviewers were helpful and gave keen insights on my answers.
Data Modelling Round 1 hour
This round was basically based on the creation of schemas for the fact and the derived tables based on the business usecase. The usecase was related to movie streaming and review generation. I had to create the derieved tables and write the sql queries for them. This was by far the most difficult round. I was able to pass this round by an inch. This too was taken by a staff and a principal architect.
DSA ROUND 1 hour
This round was a complete DSA round but the catch was that this was in python only. Since I code in JAVA so I asked them about to use JAVA. They helped me with the python syntax. The expectation was to solve 3 questions on the binary search tree and pass the testcases provided. The round went well.
Data Quality 1 hour
This round was a creative one. I was provided a pdf which consisted of 10 tables with data. I had to go through them in 10mins and find what can be the possible bottlenecks and the error if we persist these data in the production database and datawarehouse. I was expected to find the errors and give possible solutions and write storing rules of data population. This was by far the best round for me.
The recruiter informed me in 2 days that my reviews were very good and I was selected for the position.
Offer Evaluation: https://leetcode.com/discuss/post/6763608/indeed-l2data-engineering-offer-evaluati-4lb6/
Interview Questions (7)
Given a list of recipes and available raw materials, determine which dishes can be prepared. This was a DSA question based on hashmap and strings, considered a LeetCode medium level problem.
A Data Structures & Algorithms question utilizing Breadth-First Search (BFS), conceptually similar to the LeetCode 'Rotten Oranges' problem.
A Data Structures & Algorithms question based on the LeetCode 'Course Schedule II' problem, with a minor modification.
Discussion revolved around my resume, in-depth Spark applications, proud projects, how I implemented them, potential improvements, and deep analysis of various databases and their use cases.
Design a high-level data pipeline where one source is a database table and the other is an API.
Design schemas for fact and derived tables for a business use case involving movie streaming and review generation. This included creating derived tables and writing corresponding SQL queries.
Given a PDF with 10 tables of data, identify possible bottlenecks and errors if the data were persisted in a production database and data warehouse. Provide solutions and define storing rules for data population.
Summary
I had a fantastic interview experience with Indeed in Sept-October 2022 for an SDE2 role in Hyderabad, successfully navigating multiple DSA, System Design, Code Review, and Behavioral rounds to receive an offer.
Full Experience
Phone Screen
My interview process began with a phone screen, which I found to be a fantastic experience, one of the best among all the firms I interviewed with. For the first 20 minutes, we discussed system design and architecture concepts. Following that, I worked on Data Structures and Algorithms (DSA) questions. I managed to fully solve two DSA problems and started on a third before my time ran out. The expectation in this round is to solve as many questions as possible within the 30-35 minute DSA slot. I was informed that solving one question fully and starting a second is generally enough to clear this K*r*t round.
After two days, the recruiter informed me that I had successfully cleared the phone screen, and they would proceed with scheduling the next set of interviews, which at Indeed, are taken in two phases.
Phase 1 Onsites
Round 1 (DSA)
This round focused on a question about efficiently storing a binary tree in memory. The interviewer challenged me to optimize the space typically used by a struct TreeNode, which stores both left and right pointers along with the value. I proposed representing the binary tree as an array, and our discussion revolved around this approach, exploring its applicability and implications for both complete and sparse trees.
Round 2 (System Design)
I was tasked with designing a company review system for Indeed. This is a commonly asked system design question, and I believe I performed quite well. The interviewer was very supportive and helped guide me through the discussion, covering requirements, various services, API contracts, database choices, and the tradeoffs inherent in my design decisions.
The recruiter contacted me two days later to confirm I had cleared Phase 1 of the onsite rounds and would be moving forward to the second phase.
Phase 2 Onsites
Round 1 (DSA)
This round presented a truly HARD question, which was a modified version of a hard-level LeetCode problem. The problem statement was: "Given m sorted lists with each having an average size of n (no duplicates within the same list), find out the most frequently occurring k elements and return them sorted by the frequency. Note: Duplicates don't occur in the same list, but they can surely come up across the lists." An example provided was for m=4, k=2 with lists [1, 2, 4, 8], [1, 2, 3], [1, 3, 5, 11, 12, 16], and [1, 2], where the expected answer was [1, 2].
My initial, brute-force approach involved iterating over all lists and storing element frequencies in a map, resulting in O(m * n) time and O(m * n) space complexity. The interviewer then immediately posed a follow-up: what if the lists were extremely large, so much so that they couldn't all fit on a single node? This required me to optimize the space complexity. I came up with an optimal solution using two heaps, achieving O(m + k) space complexity, which I then coded within the allotted time. We also discussed a few follow-up scenarios related to this solution.
Round 2 (Code Review)
This round is unique to Indeed. I was given access to a link displaying a side-by-side diff of a Pull Request containing a couple of files. My task was to review the PR and provide comments as I would during a normal code review. I selected Java as my language. I felt I performed fairly well, though honestly, I'm not exceptionally proficient in Java, despite using it at work. This round went okay-ish.
Round 3 (Resume Deep-Dive)
This round was conducted by a senior engineering manager. They asked me to choose one of the projects I had worked on and explain it in detail. I had extensively prepared for such rounds, particularly those focusing on leadership principles, so this discussion went smoothly for me.
Round 4 (Closing Round)
My final round was with a Director of Engineering. They asked some basic behavioral questions and inquired about my overall interview process experience. We also briefly discussed the team they head. It was a 30-minute round, and I genuinely liked both the director and the team.
Two days later, the recruiter verbally informed me that the response was positive and an offer would be extended. I shared the compensation details of another offer I had, and it took about 10 more days for the official offer letter to be rolled out.
Thankfully, the offer arrived just before Indeed announced a hiring freeze, and they specifically assured me that no official offers would be rescinded. I accepted the offer the moment it came out, as the recruiter had indicated there was pretty much no scope for negotiation, and I was also quite tired of giving interviews. I verbally accepted the offer then. As of today, my background verification is complete, and I am excited to be joining Indeed this month.
Interview Questions (6)
The interviewer asked me to optimize the space used when storing a binary tree in memory. Instead of the typical struct TreeNode with explicit left and right pointers and a value, the goal was to find a more space-efficient representation. The discussion then explored how this approach would apply to both complete and sparse trees.
I was tasked with designing a company review system for Indeed. The discussion covered various aspects including requirements gathering, identifying different services, defining API contracts, choosing appropriate databases, and analyzing the tradeoffs of different design decisions.
Given m sorted lists, each with an average size of n (no duplicates within the same list), the problem was to find the k most frequently occurring elements and return them sorted by frequency. Duplicates were allowed across different lists but not within the same list. An example provided was for m=4, k=2 with lists: [1, 2, 4, 8], [1, 2, 3], [1, 3, 5, 11, 12, 16], [1, 2], where the expected answer was [1, 2]. A crucial follow-up involved optimizing space complexity if the lists were extremely large and couldn't fit into a single node's memory.
This round, unique to Indeed, involved reviewing a Pull Request. I was given access to a link showing a side-by-side diff of a PR with a couple of files. My task was to provide comments as I would typically do during a professional code review, using Java as the chosen language.
A senior engineering manager asked me to select one of the projects from my resume and explain it comprehensively, covering its design, implementation details, challenges, and impact.
During the closing round with a Director of Engineering, I was asked some fundamental behavioral questions. Additionally, the director inquired about my overall experience with the interview process.
Preparation Tips
My preparation involved extensive practice, particularly for Data Structures and Algorithms, and thorough preparation for behavioral questions, especially those aligned with leadership principles. I believe this comprehensive approach was key to my success.
Summary
I interviewed for a Software Engineer role at Indeed in Hyderabad in June 2022. Despite successfully clearing the initial Phase 1, I unfortunately received a rejection after completing the comprehensive Phase 2 interviews.
Full Experience
My journey with Indeed began when a recruiter contacted me via LinkedIn regarding an opening for a Software Engineer position. I am currently an SWE2 at a Fintech company with 3 years of experience.
Karat Round (1 hour)
This round started with a 10-minute session where I was given a choice of theoretical questions, and a few were asked from there. Following this, the interviewer presented multiple Data Structures and Algorithms (DSA) questions until the hour was exhausted. I was given three questions, all of medium difficulty.Phase 1: (Cleared)
This phase consisted of two rounds. Clearing these would lead to an invitation for Phase 2.- Round 1: High-Level Design (1 hour): I was asked to design the High-Level Design (HLD) for a new review requirement on the Indeed site.
- Round 2: Whiteboard 1 (1 hour): I encountered two DSA questions, which were described as being of LeetCode medium difficulty and sort of design questions.
After two days, the recruiter informed me that I had cleared Phase 1 and wished to schedule Phase 2, which would happen on the same day.
Phase 2: (Rejected)
Four rounds were scheduled for Phase 2, all on the same day.- Round 1: Whiteboard 2 (1 hour): This round challenged me to design a compiler for Python, specifically focusing on handling basic functionalities like indentations and colons after statements.
- Round 2: Resume Deep Dive (1 hour): The interviewer asked me to explain one of the projects detailed on my resume. Subsequently, I was asked how I would accommodate a new functionality they provided into that project.
- Round 3: Code Review (1 hour): This interview began with a discussion on the process of code reviews and what specific aspects I typically look for. Following this, I was given a sample code on Hackerrank and instructed to review it by adding comments, without the need to explain my comments.
- Round 4: Close Interview (30 minutes): A senior engineering manager joined this round. He inquired about my interview experience and suggested ways to improve the process. He also asked me to rate my performance in each round and posed several behavioral questions.
Four days after completing Phase 2, I received an email stating that I was rejected.
Interview Questions (4)
I was asked to design the High-Level Design (HLD) for a new feature on the Indeed site that allows users to leave reviews.
I was asked to design a compiler for Python that could handle basic functionalities such as indentations and colons after statements.
I was asked to explain one of the projects from my resume in depth. Following this, the interviewer posed a scenario asking how I would accommodate a new functionality he provided into that project.
The interview began with questions about my approach to code reviews, including the process I follow and the specific aspects I typically look for. Afterwards, I was given a sample code on Hackerrank and instructed to review it by adding comments, without needing to explain my comments.