Four rounds at Amazon's Seattle HQ, then nothing
| Company | Amazon (AWS) |
|---|---|
| Team | S3 Glacier |
| Role | SDE1 |
| Location | Seattle HQ, onsite |
| Date | August 2026 |
| How it started | Cold application |
| Rounds | 4 |
| Outcome | Rejected. No feedback. |
That was my fifth rejection from Amazon or AWS in three years at interview stages.
The loop
Four rounds, each an hour, each split the same way: behavioural first, then code. The last column is my guess at what the interviewer wrote in the debrief, on the scale they actually use, not how I felt about it.
| Round | First 15 min | Next 45 min | Self Rating |
|---|---|---|---|
| 1 · Bar Raiser | Leadership Principles, the full hour | No coding | Strong Hire / Hire |
| 2 · Technical Coding | GenAI experience, and how I use AI tools day to day | Sliding-window substring problem | Hire |
| 3 · Technical Coding | Leadership Principles | Parenthesis matching, stacks | Lean Hire |
| 4 · LLD | Leadership Principles | Rate limiter, low-level design in C++ | Hire |
On both coding rounds I gave brute force, then a better approach, then the optimal one.
Round 3 is the one that went wrong, and not because of the problem. The interviewer was unresponsive to the point where I could not tell whether anything I said had landed. I explained the approach three ways. I got almost nothing back. You cannot course-correct against silence, and by the end I had no idea whether I was being evaluated or just waited out.
Then nothing
No feedback. No call. The recruiter simply stopped responding.
Candidates are told to be thorough and to have high standards, and then the process does not hold itself to the standard it interviews for.
What I actually think, after 5 rejections
The loop did not measure very much. Two coding rounds where I produced brute force, better, and optimal, and one of them still ended with an interviewer who could not follow the explanation. That is not a signal about my ability and it is not really a signal about theirs.
Five rejections from one company over three years is either a strong signal about me or a weak signal repeated five times. I genuinely do not know which, and I think anyone who tells you they know which for their own case is guessing.
What I would do differently
I gave my best shot and could not have done better. I will not change anything about my preparation. I will continue to solve more problems and hope for luck to favor someday.
🥹 If this helped, a like means a lot.