Amazon SDE2 Interview Experience
Summary
I interviewed for an SDE2 role at Amazon, going through multiple rounds including LLD, DSA, HLD, and a Bar Raiser round. I was ultimately downleveled to SDE1 due to mixed feedback in the HLD round.
Full Experience
Started with OA, solved both the questions and got call for interview. All rounds happened on Chime. Each round had 2-3 leadership-based question.
Round 1: LLD LP - 30 mins same LLD question from here - https://leetcode.com/discuss/post/6708128/amazon-sde2-hyderabad-interview-experien-j65b/ Feedback - Good
Round 2: DSA LP and Resume - 35-40 mins Median of a stream of integers. Gave optimal solution. Feedback - Good
Round 3: HLD LP - 30 mins Its a variant of reservation system, where a user can book a slot. and there is maximum limit for bookings for each slot. Discussed more about concurrency, multiple users selected a particular slot but has not booked. Feedback - Mixed
Round 4: Bar Raiser Interviewer joined 5-10 mins late LP - 20 mins This is a variant of toposort. Requirements were not clear. Initially misunderstood the question and explained approach using BFS and wrote code. Interviewer highlighted this during dry run. Explained DFS approach after that, was not asked to code it. Closed the interview 5 mins early
Downlevelled to SDE1 because of the mixed reviews in HLD round.
Interview Questions (4)
Low-Level Design Question (Refer to Link)
The LLD question was identical to one discussed in another post. Please refer to the provided link for details.
Median of a Stream of Integers
Design a data structure that supports adding new integers and finding the median of all elements received so far.
High-Level Design: Reservation System with Concurrency
Design a reservation system where users can book slots. Implement a maximum limit for bookings per slot and discuss handling concurrency issues, specifically when multiple users select a slot but haven't completed the booking.
Topological Sort Variant
A variant of a topological sort problem. Requirements were initially unclear, leading to an initial incorrect approach using BFS. After correction during a dry run, a DFS approach was explained.