Amazon Software Development Engineer (SDE) Interview Guide

The Amazon Software Engineer interview process typically runs in this order:

• Recruiter call (can come before or after the online assessment)
• Online assessment (OA), typically only for SDE I and SDE II candidates 
• Technical phone screen
• Onsite loop of 4-5 interviews, each w/ a technical and Leadership Principles portion:
  Coding
  Gen-AI Fluency round (in some 2026 loops)
  Design
  The Bar Raiser interview
• Debrief and team matching 

The entire process is said to take 3-6 weeks but you can expect news on whether you move to the onsite in as little as 2 (business) days. Of course, there have been statistical outliers to this timeline, with tail-end cases taking several months.

If you’re interviewing to be an EM, please switch to our Amazon Software Developer Manager interview guide.

Amazon structures its entire operational culture around its 16 core Leadership Principles (LPs). Each interviewer that you'll meet in your loop will be explicitly assigned to evaluate about two or three of these principles using a targeted STAR (Situation, Task, Action, Result) format probing.

To give you a better edge, we recommend that you take the classic STAR framework and build it out into STAR+ (Situation, Task, Action, Result + the X-Factor: Learning) to structure your stories. The X-Factor at the end - Learning - is your definitive chance to demonstrate intellectual humility, and helps you reflect on a personal takeaway on how that situation changed you for the better. 

For SWE candidates, the most heavily probed LPs are Ownership, Bias for Action, Dive Deep, Earn Trust, and Have Backbone; Disagree and Commit.

To understand the line of questioning and the recommended approach you should take as a promising candidate, please read our article: Preparing for Amazon’s behavioral interview.

This is a 30-45 minute phone/video conversation with a recruiter or sometimes with the HM. Amazon clarifies that this is a formal interview that’ll screen and understand your basic qualifications and technical background. It can also include a couple of behavioral questions that are tied to Amazon’s 16 LPs and some scenario-based questions (invented or not) that are specific to the SDE role. Comp also comes up here. You’ll have time at the end for your own questions about the role/ team. 

Prepare to answer questions like:

• Tell me about yourself. Why are you applying to Amazon? 
• What are you currently working on day-to-day? 
• On a scale of 1-10, how would you rate yourself in [language framework]?

From this point on, the recruiter is your main POC so you should take the time to: 

• Clarify which subjects/ skills your interviews will focus on
• Ask which level you’re being considered for so you know how to shape your prep

The OA is a HackerRank test of ~2 hours that you take on your own within 7 days of receiving the link. 

The sections you’ll face depend on your level:

• SDE II:
  a 90-min coding challenge (timed)
  15-20 min of system design scenarios
  a work-style survey of about 8-10 minutes.
• Full-time student and new grad roles: 
  70 min coding assessment  
  15 min work-style assessment
  60 min work simulation
• SDE internships: coding assessment + work-style assessment that takes 90 minutes in all.

Amazon's SDE II OA page and university OA page describe this assessment in detail, including rules and what references you’re allowed to look up. 

The coding portion has 2 versions:

• Standard: You’ll solve 2 algorithm problems in HackerRank's editor, in the language of your choice. The code has to compile and run on every question for you to move forward. 
• AI-assisted: You’ll face one algorithm problem and a task inside a code repository. Prepfully candidates say that these tasks took them 40 and 60 minutes respectively. 

You get to pick the framework for the repo when you get started but you can’t switch halfway through. Candidates report the following options:

• Node.js (MERN)
• Python with Django
• Java with Spring Boot

The repository opens in a browser-based version of VS Code, where the README describes the tasks. Each task has predefined tests that you run from the IDE. An AI assistant is provided and using it is optional. You have to remember that the assistant won't (can’t) write a full working solution. Also your conversation with the AI is shown to your evaluator but doesn't affect your score.

HackerRank's guide to the Amazon OA covers the process and rules for both editors in detail.

For the system design portion, you’ll apply your skills to fictional situations that involve simulated colleagues and stakeholders. The scenarios draw on design concepts such as designing for maintainability and weighing trade-offs and you’ll have to rate which action you’d recommend and which one you wouldn't. You don’t draw and diagram here. 

The work simulation portion is self-explanatory. A workday will be simulated in which you have to answer emails and questions from virtual coworkers and handle the stuff an Amazon SDE would. 

The work-style survey is a direct test of your alignment to their LPs via multiple choice questions about how you like to work. From a given set of statements, you’ll have to choose which one’s most like you. 

Amazon's Online Assessment Practice Test is open to anyone. There’s exactly zero risk to practicing, so please go ahead and make good use of the practice link to stay ahead of the curve.

The technical phone screen is a 60-minute video interview with an Amazon engineer/senior leader. Time is split evenly between coding questions and behavioral questions related to the LPs. 

Coding-wise, you’re given one main problem (or 2 short ones) of medium difficulty focusing on standard DSA. Expect the interviewer to add follow-ups that change the constraints or introduce new data. You’ll have to write clean, bug-free code in the shared editor without the safety net of autocomplete. 

Expect questions like:

• Find the maximum value in every window of size k as it slides across an array.
• Given a 2D grid of 1s and 0s, where 1 is land and 0 is water, count the number of islands.
• Design a data structure for an LRU cache where both get and put run in O(1) time.

It’s super important to establish all constraint boundaries by asking clarifying questions before jumping in to solve a question. Verbalize your brute-force v/s optimal approach and time/space complexity before you start writing code or before the interviewer has to bring it up. 

Also take some time to practice tracing test cases manually. It takes a bit of getting used to, especially given the modern-day engineer’s habits. You don’t want to find out you’ve lost touch with dry-running in the interview. 

The onsite loop features a set of 4-5 interviews that are 60 minutes each. These interviews will be run by people from the hiring team and one Bar Raiser who’s outside it. We’ll address the implications of that in the Bar Raiser section. 

You’ll typically face:

• Coding
• Gen-AI Fluency round (in some 2026 loops)
• Design
• The Bar Raiser interview

Every interview will accommodate technical questions and 2-3 behavioral, LP-focused ones. To keep the guide lean, we’ll stop referencing the behavioral portion in every round and stick to the technical portion alone.

This round is a live coding round conducted on a shared editor. You’ll face 1-2 problems that are typically LeetCode medium/hard. Interviewers state the problem without delving into every detail. This is an opportunity for you to ask targeted clarifying questions and for them to later change requirements or constraints to see how well you accommodate them into your solution. 

Amazon says your coding should focus on efficiency, reliability, robustness, portability, maintainability and readability.

For SDE II, Amazon says candidates should prepare for 3 separate coding interviews, each built around one of the following competencies:

• Data structures and algorithms: how they work, what their operations cost in time and memory, when to choose one algorithm over the other, etc.
• Problem solving: how you break down ambiguity by uncovering requirements, building to suit those and then extending to incorporate new ones the interviewer adds
• Logical and maintainable code: how readable and maintainable your code is so someone who inherits it isn’t like a deer in headlights. Importance is given to clear separation of concerns across methods and classes and test names that accurately describe the requirements of each test. 

Expect questions like:

• Build a cache for a file-system store that has an eviction policy. How would you change it to hold file contents in place of file paths?
• What's the length of the longest strictly increasing subsequence in which consecutive elements differ by at most k?
• Given a time in HH:MM format, what's the next time that reads as a palindrome? 
• Given a set of binary strings, what's the largest subset you can pick that uses at most m ones and at most n zeros in total?

Aptly named, this round is built to understand how you use generative AI tools in your engineering work. Many versions of it also include a coding problem. When it does, it runs exactly like the coding round described earlier. This round mostly shows up for SDE I and new grad/internship loops. 

Most candidates describe the questions to be practical with nothing to do with how the models work/ other technical aspects. Some interviewers even use hypothetical scenarios like a critical production issue at YouTube that had to be fixed within an hour using AI tools. 

Even if the round isn’t scheduled as a stand-alone, you’ll likely face these questions elsewhere in your loop:

• How do you use AI to increase your engineering productivity?
• How do you decide if you can trust an AI? How do you validate its output?
• Tell me about a project where you used AI to solve a problem.
• Has AI ever messed up your work? Did you continue using AI for that task?

While Amazon is yet to officially publish how the round is scored, our coaches say right now, the round draws a line at “depth of use”. 

In design rounds, you either go for something low-level or high-level. You’ll be judged on practicality, accuracy, efficiency, reliability, optimization and scalability. 

SDE II and SDE III candidates should expect at least one system design round and for SDE I candidates, design comes up as a low-level discussion inside coding rounds itself. 

In the low-level version, you’ll identify the entities in the problem and define the classes and the relationships between them. Then, you’ll write code for the core classes. Once your base design works, the interviewer will add a few requirements to see how well it extends.

Questions here will usually read like:

• Design and implement a meeting room scheduler 
• Design a vending machine leasing system with clean interfaces for the machine's states and for payments 

In a high-level design round, you’ll start by gathering requirements and move on to sketching the API and the data model. After that, you’ll choose how to store the data and work through how the system scales and how it handles failures. In 2026 reports, HMs also asked product questions about the design as follow-ups.

Candidates have reported the following questions:

• Design a real-time error log monitoring system that handles more than 1 million writes a second and alerts within a second
• Design a Facebook-style news feed
• Design a music streaming service like Spotify
• Design a device backup scheduler that saves a device's settings and media and restores them onto a new device 

Since the system design format is standardized across the industry, we covered all of it in our System Design guide if you’re looking to brush up on the format or looking for tips.

This is an hour-long interview with a trained interviewer that’s outside your hiring team. Their role is to ensure that every new hire performs better than 50% of current employees in a similar role. 

In Memoirs of an Amazon Bar Raiser, Carlos Arguelles writes “BRs are a unique part of the Amazon culture. You lead the debrief (which includes all the interviewers in the same room debating their data points) and make the final hire/no hire decision. You have veto power over all hiring decisions (a hire cannot happen without your thumbs up.)”

What happens in the round varies. Many Bar Raiser interviews are mostly behavioral and built around a deep dive/retro on one of your previous projects. They’ll ask about the extent of your contribution or ask you to defend certain decisions you made. 

Other versions follow a similar process but include a short coding/design problem. 

Generally speaking, you should be prepared for this round to be a rigorous LP-alignment audit with some pressure-testing around your narrative. 

Read our Amazon's Bar Raiser Interview: A Comprehensive Guide to the Process for 6 tips on how to answer their questions.

Amazon grades candidates by rating specific leadership principles as either a Strength (ST), Satisfies (SA), Development Needed (DN), or No Evidence (NE). Interviewers use these ratings to then cast a final vote to hire you on a scale of Strongly Inclined > Inclined > Not Inclined > Strongly Not Inclined.

New grads have historically been placed on a team just before their starting date, with little say in which one.

• SDE I: $182K
• SDE II: $274K
• SDE III: $387K
• Principal SDE: $641K
• Senior Principal SDE: $1.04M

View Amazon SDE compensation data on Levels.fyi