The Amazon SDE Intern Interview Process

Most descriptions of Amazon hiring describe the full-time Loop, which is not what an internship candidate goes through. This section covers the intern version, which has fewer stages and shifts the weight toward the front.

• Application and resume screen. University recruiting runs on a season, so the calendar matters more than at other times. Applications commonly open in late summer for the following year.
• The online assessment. A timed, remotely proctored assessment sent by email. It typically mixes coding problems with a work-simulation exercise, where you respond to workplace situations, and a work-style survey aligned to the Leadership Principles.
• Interviews. Usually one or two video interviews on Amazon Chime, each around 45 to 60 minutes, mixing a coding problem with behavioral questions.
• Decision and team match. Offers may arrive before a specific team is assigned, with matching happening afterwards.

Timelines vary widely by season and volume. Candidates report anything from two weeks to a couple of months between the assessment and an interview invitation.

Three differences change how you should prepare. The obvious one is length: a new-grad SDE 1 candidate faces a Loop of four or five interviews, while an intern is more often looking at one or two. The stage-by-stage new-grad version sits under /answers/topic/company-interviews.

The second is where the filtering happens. With fewer interviews, the assessment does more of the work, so a weak assessment is harder to recover from than it would be later in a longer loop. Treat it as a real round rather than a formality.

The third is scope. System design does not appear at intern level, and the coding problems stay closer to data structures, strings, and graphs than to distributed systems. What does not shrink is the behavioral component, which is the part interns most often underestimate.

They are scored, and interviewers write down what you say against them. The common mistake is assuming that a student has nothing to draw on, so the Leadership Principle answers get improvised while the coding gets weeks of preparation.

Coursework, a hackathon, a part-time job, a club, and a personal project are all valid sources. Consider a second-year student asked for a time she had to deliver with incomplete information. Her answer was a group assignment where a teammate stopped responding two days before submission, what she cut, and what shipped. That is a complete story with a decision and an outcome, and it scored the same way a work example would.

Prepare four or five stories with the situation, your specific action, and the result, and expect follow-ups asking what you would do differently. Practising them out loud at speaking pace is what /mock-interview is for; written notes rarely survive contact with a live follow-up.

The two stages have different answers, and conflating them is how candidates get into trouble.

The online assessment is timed and remotely proctored. Amazon's assessments monitor the session, and using outside help there is a policy violation regardless of what any tool makes technically possible. There is no honest reading of that stage in which an assistant is appropriate.

A normal video interview on your own machine is a different situation. A local assistant that transcribes meeting audio and shows suggestions in an overlay on your screen is technically workable, and no bot joins the call. It is still out of scope the moment you share your screen, the interview is recorded as a one-way response, or you are on a laptop the company manages. Those boundaries are written out at /answers/topic/detectability, and they are worth reading before you rely on anything.