Your Amazon Prep Plan

This guide runs to roughly a dozen lessons across five sections. Before you open any one of them, take the plan that sequences them. It walks you through the coding patterns, the system design, the behavioral stories about Amazon's named values, and the offer at the end. The phases are sized as a share of whatever prep time you have, so the same plan works on a six-week runway or a frantic long weekend. And one thing before anything else: at Amazon, behavioral is the centerpiece of the loop, and your plan has to spend time like it.

The Amazon angle

About a quarter of candidates who clear Amazon's technical bar still get rejected on behavioral grounds. Amazon scores its 16 Leadership Principles (the company's named values, like Customer Obsession and Ownership) inside every round, coding and design included. There is no single interview where the soft stuff lives and the rest where it does not. A candidate on the technical borderline can still get a hire when the principle stories land well. The reverse almost never happens. Weak stories sink an otherwise clean technical loop. So this plan hands behavioral a far bigger slice than a generic FAANG plan would, and front-loads it instead of saving it for last.

How to spend your prep time

Each phase has one focus, a rough share of your time, and a checkpoint. The percentages assume you are targeting L5, the SDE II band where most experienced hires land. Below L5, drop the system design phase entirely and pour that share into coding and your principle stories. Do not move past a checkpoint until you can clear it on demand. A shaky base just collapses once the follow-up questions start and you are coding in Amazon's editor, which never runs your code for you.

Why behavioral gets the biggest slice

Look at the shares again. Coding and behavioral come out roughly even, and system design trails both. That split is deliberate. A generic prep plan would dump most of the time into LeetCode, on the theory that the technical bar is the hard part. Amazon's loop spends a fixed chunk of every round on principle stories, and plenty of candidates who code well still wash out there.

The Amazon angle

The trained interviewer who can sink you here is the Bar Raiser, someone assigned from outside the hiring team whose entire job is to protect the bar. They hold a veto over the hire, and they lead the debrief where the room argues your case, with their focus fixed on culture and the Leadership Principles. A tidy binary search barely registers with them. So when this plan gives stories a full quarter of your time, it is buying down the exact risk the Bar Raiser exists to catch. They have seen plenty of strong coders who never gave a single concrete example of the judgment and ownership Amazon promotes on, and those candidates get no-hired. Eight to ten stories, each one quantified and each flexible across several principles, is the asset that survives their probing. Build it early.

In the room

The habits below have nothing to do with knowing more. They are about showing what you already know in the way Amazon scores it, round after round. That is often what splits a prepared candidate from a merely smart one. Treat them as muscle memory.

• Clarify before you code. Restate the problem and pin the edge cases before writing a line. Amazon's editor will not run your code, so the interviewer is grading your reasoning, not your compiler.
• Think out loud the whole time. Each interviewer writes up your round afterward and submits a hire or no-hire vote before the debrief. Silent brilliance gives them nothing to quote in your favor.
• Say "I" not "we" in every story. Principle answers are scored on what you personally did, roughly 80% on your own Action and Result. "We shipped it" reads as hiding.
• Quantify the result. The single most common piece of critical feedback at Amazon is "needed more metrics." Cut latency by how much, saved how many dollars, unblocked how many people. Numbers turn a story into evidence.
• Test your code by hand before you call it done. Trace a real input end to end. With no run button, this is the only proof the interviewer gets that it works.
• Treat the recruiter as your ally. Ask what level you are being assessed at and what the timeline is. Level is set by loop performance and is hard to move later, so knowing the target shapes how you prepare.

The Amazon angle

"Have Backbone; Disagree and Commit" hides a trap, and it is the principle candidates most often misread. Interviewers want a story where you respectfully pushed back with data, then fully committed once the call went the other way. They do not want a war story about proving you were right all along. If your conflict story has no "and then I got behind the decision and made it work" ending, it scores against you. Build that beat in on purpose.