Flipkart | SDE2 | Bangalore | April 2024 [Offer]

My interview journey for the SDE-2 role at Flipkart began in the first week of March 2024, culminating in an offer in April. I have 2.6 years of experience, and the entire process consisted of five rounds.

Round 1: Machine Coding/LLD (90 minutes)

This round was split into 30 minutes for problem statement discussion and 1 hour for coding. The task was to design an in-memory vehicle rental service, similar to Zoomcar, specifically for a single city with multiple branches. Key functional requirements included:

• Managing multiple branches, each with a limited number of different vehicle types.
• Booking vehicles for predefined 1-hour slots on a single day (00-59).
• Fixed pricing per vehicle.
• Features for onboarding new branches and vehicles.
• Renting a vehicle for a time slot and type (defaulting to the lowest price).
• Displaying available vehicles for a given branch.

The system had to be CLI-driven and utilize in-memory data structures, prohibiting any database usage. I was given 1 hour to code. My approach involved analyzing requirements to identify entities, then creating Model/POJO and DTO classes. I developed services and controllers for user interactions, prioritizing loose coupling and dependency injection. For persistence, I relied on Java Collections, ensuring thread safety (though optional). I implemented a minimal exception hierarchy and global exception handler. To optimize operations like vehicle booking (Requirement 3), I considered using a min-heap, accepting an overhead on vehicle insertion queries. For displaying available vehicles (Requirement 4), I overrode the toString method. I also applied design patterns like Factory (for vehicle generation), Strategy (for pricing), and Singleton (for managing thread-safe instances). Post-coding, I demoed the system with various use cases and answered questions on design patterns used, schema design, and database storage/query patterns.

Round 2: DSA/PSDS (60 minutes)

This round involved two LeetCode Medium problems.

1. The first problem was an overlapping interval scenario, which I approached using a line sweep algorithm.
2. The second problem was a standard tree question: "Print Level order traversal of a binary tree using Given Inorder and Preorder Traversal". I was required to provide a clean, compilable solution and explain my intuition clearly. I leveraged concepts like Sorting, TreeSet, Binary Trees, and Recursive Subproblems. Having solved around 120 medium problems on LeetCode, I found this round manageable.

Round 3: Design (LLD+HLD) (60 minutes)

I was asked to design BookMyShow. I began by gathering functional requirements and mapping out a user journey, which included searching for shows, selecting screenings and seats, proceeding to checkout, completing payments, and viewing tickets. I then designed APIs for each step, such as GET /shows/queryKeyword, GET /shows/selectSeats, POST /shows/confirm (which would acquire soft row locks), POST /shows/checkout, and GET /bookings/mybookings. I detailed how different objects would interact with these APIs, and how operations would be handled internally down to the database level. Critical discussions included managing concurrent seat bookings, failure scenarios, data store choices, race conditions, and resource contention. For the High-Level Design (HLD) part, I crunched numbers for user base, bandwidth, traffic, and storage requirements. I also discussed Non-Functional Requirements (NFRs), sharding strategies, load balancing, and caching mechanisms.

Round 4: Techno-Managerial Round (60 minutes)

This round was with a Senior Engineering Manager and focused on my profile and past experience. I was asked to deep-dive into a specific project, discussing its problem statement, technical solution, business case, coding, and deployment. We explored alternative solutions, potential improvements, and enhancements. Since it was a microservices project, I explained its entire ecosystem. Questions also covered NFRs, high-level architecture, DevOps practices, pipelines, and deployment choices. My mention of Redis led to questions on Redis clustering and replication. Additionally, I faced scenario-based technical and behavioral questions, focusing on incident/bug handling, downtime issues, and ownership. The interviewer made me comfortable, allowing me to articulate my thoughts freely. The round concluded with a discussion about my career progression and role development.

Round 5: HM Connect

This was a casual discussion with the Hiring Manager, focusing on the technical ecosystem, Lines of Business (LOBs), team structure, and culture fitment. I asked several follow-up questions, and the manager provided clear insights into the organization's scale and engineering culture. We also touched upon individual progression and benefits.