Ola Interviews

Round 1: Online Assessment

I started with an online assessment on the Cocubes platform. It was a 40-minute round with two medium-level questions. I managed to code both problems within 30 minutes. However, the platform itself was quite frustrating, offering only two test cases to verify my code.

Round 2: System Design / HLD/LLD

This round focused on High-Level Design (HLD) and Low-Level Design (LLD, though it leaned heavily into HLD). The interviewer also asked some questions about my previous projects before diving into the core design problem. I was asked to design an online streaming platform.

I began by discussing the necessary features and services, outlining the functional requirements (FRs), non-functional requirements (NFRs), and network estimations. Following that, I opened ExcelDraw to design the components of the application. For the first 15 minutes, the interview seemed to be going well, with the interviewer in sync with my approach.

However, after this initial period, the interviewer completely lost focus. They seemed to be doing their own work during my interview time and barely paid attention, often requiring me to call out to them twice to get their attention for my questions or points. I genuinely felt bad for wasting my energy trying to drive the entire interview while they were disengaged. Despite this, I gave my absolute best to design the streaming platform, covering all possible components. This truly was my worst interview experience to date.

I applied for the SDE 1 role at OLA through a referral. My background includes a B.Tech in Computer Science from a Tier 3 institution, with 1 year of prior experience at a startup.

Online Test on Moocha

I received the online test link and Round 2 invitation simultaneously. The test lasted 60 minutes and consisted of:

• SQL (5 mins): An easy SQL question based on joins, where I had to find the max value from a given schema.
• 3 Coding questions (55 mins): These were of easy-medium difficulty. One was based on LIS, another involved easy array manipulation, and the third was a DP-based question. I managed to solve the first two questions completely. A minor issue was that the platform only supported Java and Python, whereas I usually code in C++.

Round 2 - PSDS (Problem-solving and Data Structure)

This round was 45 minutes long, and I was asked three questions:

1. Shuffle Songs Without Repetition: I was asked how I would play a list of songs randomly such that no song repeats. My initial thought was a brute-force solution using random number generation. For the follow-up, "Can you do it in constant space?", after receiving some hints, I came up with a solution:

   for i from n to 1
     r = random number from (1 to i)
     play the song
     swap the song with ith song
   

   I was asked to code it, but I forgot the exact C++ random syntax, so I used a pseudo-random generator.
2. Consecutive Numbers Summing to N: The problem was to find sets of consecutive numbers summing up to a target n, for example, [2,3,4] and [4,5] for n=9. I began explaining a solution using the formula Sum = n*(n+1)/2, but the interviewer asked for a brute-force approach first. I explained the brute force and then offered a few optimizations.
3. Area Under Histogram Variation: This was a variation of the famous LeetCode "Largest Rectangle in Histogram" problem. I started with a brute-force explanation. The interviewer then prompted me to optimize using a different data structure, so I explained the stack-based solution. When asked how I identified the use of a stack, I elaborated on the concept of a monotonic stack. I was not required to code this problem.

Towards the end of this round, I was also asked "Why do I want to leave my current organization?"

Round 3 - Design Round

In this round, I was asked to design a movie ticket booking site/app, similar to BookMyShow. The discussion was very high-level, focusing on my approach and answering follow-up questions. No coding was involved.

Round 4 - Hiring Manager Round

During this round, two interviewers accidentally joined at the same time, which initially confused me. However, one interviewer was very friendly, introduced himself, and helped me feel comfortable.

• I was asked to introduce myself and discuss my projects. The interviewer reviewed my resume, and then the rest of the interview involved me explaining my projects and previous work in detail.
• Since I remembered every detail of my projects, including class names, the interviewer seemed impressed. He then shifted to asking about my other interests and hobbies.
• With about 15 minutes remaining, I was given a problem that was a variation of the DNF sort. The interviewer asked about the rationale behind using three pointers and the core idea of DNF sort.

I received an email today stating that they are moving forward with my candidature, and I am currently awaiting an official offer. I was also curious about the work culture and work-life balance at OLA, as I'm leaving my current company due to a poor work-life balance.

Outcome: I ultimately declined the offer as I received a better one.

Interviews Breakdown

I was initially contacted by a recruiter from Ola who was incredibly flexible with scheduling all the interview rounds, providing excellent support throughout the process.

Problem Solving Round 1

This round consisted of two easy Data Structures questions, primarily related to arrays and strings. Unfortunately, since it was about a month ago, I cannot recall the exact problem statements, but they were comparable to LeetCode easy problems.

Problem Solving Round 2

This round, which was 15 days prior, was scheduled for one hour but concluded in about 40 minutes after I addressed both questions successfully.

• Question 1: Given an array, I had to determine if it could be sorted by reversing any single subarray within it. For example, in [1, 2, 3, 4, 8, 7, 6, 9], reversing the subarray from index 4 to 6 would result in a sorted array.
• Question 2: This was a variation of the Two Sum problem. I needed to find pairs of numbers in the array [1, 2, 3, 4, 8, 7, 6, 0] that sum up to 6. It was assumed that numbers were positive and distinct, with expected output like [[2,4], [6,0]].

System Design Round

This round took place three days before the offer. It began with a discussion about my current project, where I was asked to draw a block diagram of its various components. Following that, the main task was to design a seat booking system similar to BookMyShow.

• GetSeatsInfo: This involved designing how to fetch seat information, including grid details, screen layouts, seat categories, and color coding for the UI. I discussed storage mechanisms in the database and the structure of request/response payloads.
• BookSeats(User u1, array of SeatsNumbers, ScreenId, TheatreId): The core challenge here was managing concurrency and consistency issues during seat booking. This discussion went very smoothly, and it felt like a collaborative effort in building the solution.

Hiring Manager Round

The hiring manager round, held the day after the design round, lasted about 1 hour and 45 minutes and was the most intense. The questions were spontaneous, and the interviewer delved deeply into every detail I mentioned, thoroughly examining my resume.

• I was asked about the most challenging project I've worked on, including the specific challenges, how I overcame them, and the tech stack used.
• The interviewer inquired about the rationale behind choosing that particular tech stack and, hypothetically, if I were to rebuild the project from scratch, what new tech stack I'd select and why.
• Since I mentioned microservices, I had to justify that choice, describe how microservices would be implemented in my current project, and specify the tech stack for each.
• We discussed flaws in my current project and the problems with NoSQL databases that might lead one to switch to relational ones.
• A complex scenario involved designing an analytics approach for my project without using a relational DB. I proposed a data pipeline and was then asked to design its workings.
• Numerous questions followed on messaging systems (RabbitMQ, Kafka, their differences, components, necessity, and real-world use cases), Redis (what it is, if it can be used as a message queue), and caching concepts.
• My experience with application monitoring tools like Kibana and Grafana was explored, including their internal workings, how I would design such a system, and appropriate database choices (especially Cassandra for write-intensive scenarios).
• Integration of logging tools, container-specific details, and log pushing mechanisms to Kibana were also discussed.
• Fundamentals of load balancing, proxy vs. reverse proxy, ZooKeeper, and sharding (when and why it's needed) were covered.
• Finally, several behavioral questions were posed, such as conflicts with managers/peers and their resolution, best and poor performance feedback received and how I handled it, and my motivation for joining Ola.
• I also had the opportunity to ask questions at the end.