Amazon SDE-2 Experience
Summary
I cleared the OA and several technical rounds but was rejected after the final hiring manager and HLD discussion.
Full Experience
Round 1: OA Cleared
Round 2: DSA
- one question related Binary Search similar to koko eating bananas
- one question related to find median from data stream
- 2 LP questions
Solved two questions
Round 3: LLD
- Design Amazon music search
- User can search by songs, artists and album
- User can create playlist
- User can play, pause and reset the song
- 2 LP questions
Solved questions
Solved using strategy + factory + state design pattern
Round 4: Bar Raiser (machine coding)
- Design Amazon distributed CRON job scheduler like (AirFlow)
- 4 LP questions
Solved Question
Round 5: Hiring manager + HLD
- 40 mins discussion on previous projects + LP
- why are you looking out for change
- weird questions like tell me about a time where you made impossible project, possible
- tell me about a time where AI couldnt generate expected code
- HLD question: Design Web Crawler within 20 mins that to in depth
Result: Rejected due to last round
Interview Questions (5)
Binary Search – Koko Eating Bananas Variant
Given a number of banana piles and a number of days, determine the minimum integer eating speed (bananas per hour) such that Koko can finish all bananas within the given days. This is the classic Koko Eating Bananas problem solved using binary search on the answer space.
Find Median from Data Stream
Design a data structure that supports inserting numbers from a data stream and retrieving the median of all inserted numbers at any point in O(log n) time per operation.
Design Amazon Music Search
Design a system that allows users to search for songs, artists, and albums, create playlists, and control playback (play, pause, reset). Discuss the high‑level components, data models, indexing strategy, and APIs required.
Design Distributed CRON Job Scheduler (Airflow‑like)
Design a distributed job scheduling system similar to Apache Airflow, capable of defining, scheduling, and executing DAG‑based workflows across multiple machines with fault tolerance and scalability.
Design Web Crawler
Design a web crawler that can efficiently traverse and index web pages. Discuss components such as URL frontier, fetching, parsing, storage, politeness policies, and scalability considerations, all within a 20‑minute discussion.