Amazon SDE 1 - Full Interview Loop Experience (Selected ✅)

Background

2 years of experience in FinTech Tier-2 college 4000 DSA problems solved over the years Used to read a lot of interview experiences during college, so posting mine here hoping it helps someone.

How it started

I came across an opening posted by a Hiring Manager on LinkedIn. Luckily had purchased LinkedIn Premium few months back, slipped into his DM, introduced myself, and applied through an internal referral link. Got a recruiter call shortly after confirming I was selected for the loop, with the first round scheduled the same day, just 4 hours later.

Round 1 - DSA + LP

Two interviewers, one main (SDE2) and one shadow. Started with LP questions around challenging projects, teamwork, and handling peer conflicts. Every answer was probed deeply with follow-up questions, nothing was taken at face value. With 35 minutes left, we moved to DSA: Design a data structure that supports Insert, Delete, and GetRandom, all in O(1). Started with brute force, then worked toward the optimized solution. Wrote clean code with time to spare and discussed tradeoffs. Cleared

Round 2 - HM Round

Surprised to get an HM round this early in the loop. The first half was a deep discussion around GenAI: Where have you used GenAI and later felt it was unnecessary? Where in your current org do you see a GenAI opportunity that hasn't been explored yet? This went on for nearly 30 minutes with a lot of cross-questions. I held my ground. Then DSA: Convert a number to its English word representation. Sounds simple, it isn't. I went down an implementation-heavy string manipulation path first, then switched to a cleaner division/modulo approach. Finished with barely a minute left, no time for follow-ups. I walked out of that round replaying every decision in my head. Two sleepless nights followed. Cleared (somehow)

Round 3 - LLD Round

Taken by an SDE2 from the hiring team. I hadn't prepared much LLD, I was mostly DSA-focused going in. Design a Parking Lot system (multiple vehicle types, different spot categories). Finished the base implementation in 15-20 minutes with clean OOP structure and called out SOLID principles throughout. Then came 5-6 follow-up requirements, and because the design was extensible, almost every one needed just a single line of code change. The interviewer noticed. LP questions covered ownership, project improvements, and handling negative feedback. This was my best round. Confidence fully restored. Cleared

Round 4 - Bar Raiser

Interviewer had 10+ years at Amazon. This one was intense from the start. Nearly 40 minutes of back-to-back LP questions, 13 to 14 of them. I went through every story I had, trying to squeeze the most out of 1.7 years of experience. Then DSA: Given a stream of numbers and a value K, print the K most frequent elements at each point in the stream. Gave brute force first, was asked to optimize. Eventually connected it to LFU Cache, tweaked the approach, and started writing DLL-based helper functions. The interviewer stopped me midway and said he understood, which I took as a good sign. Then a systems/debugging discussion: Some users are experiencing high latency in production. How do you investigate? I started with reproducing the issue, he ruled that out. Moved on to server health metrics, CPU utilization, memory, pod scaling, isolating problematic services. He kept drilling deeper on specifics. I didn't have a perfect answer, but I never stopped thinking out loud. Got a call after 2 days Verdict: Selected

What I learned

1. Never go silent. Even if you're unsure, keep talking through your thinking. Interviewers are evaluating your reasoning, not just your answer.
2. LP is not a formality. Especially in the Bar Raiser round. Prepare real, detailed stories for the top 10 Leadership Principles, with clear situations, actions, and impact.
3. One bad round doesn't mean rejection. I genuinely thought I had blown R2. Don't reject yourself before the company does.
4. LLD is fair game even for SDE1. Don't skip it. Focus on SOLID principles and writing extensible code, follow-ups become trivial when your base design is clean.
5. Network actively. A cold LinkedIn message to the HM is what started all of this. All the best to everyone preparing. Consistency is all that matters, the result is just a byproduct. Edit: Used AI to clean up the writing, because why not? Please focus on the content because everything was handwritten first and then optimized.