Amazon Applied Scientist Interview Guide

The short answer: Amazon publishes an interview prep page just for Applied Scientists. According to Amazon, you get a 60-minute technical phone screen with a senior leader on the team (some teams add a second), then a loop of four 55-minute interviews with members of its science community. The loop covers technical topics, problem solving and coding, knowledge depth and breadth, and a tech talk, and every interviewer also asks two or three Leadership Principles questions. A Bar Raiser and the hiring manager decide together, and Amazon says you hear back within 5 business days. Candidate reports describe 4 to 6 loop rounds, often with an ML system design question. Last reviewed October 2026.

Amazon applied scientist interview: quick answers

The Amazon applied scientist role

What Amazon applied scientists work on, the skills its job postings ask for, and what a week can look like. Amazon posts the role as Applied Scientist, Applied Scientist II, Applied Scientist III, Senior Applied Scientist and Principal Applied Scientist, plus Applied Science internships.

Amazon applied scientist levels and pay

Your level sets the bar you are interviewed against and the pay band of the offer.

Amazon's Applied Scientist postings do not state L levels. Applied Scientist and Applied Scientist II share a base range in Seattle-area and North Reading postings, and Applied Scientist III shares the Senior range outside New York. A candidate guide maps the titles to L4 through L7, which Amazon does not confirm on any page we read.

Pay in Amazon's US job postings

Amazon's US Applied Scientist postings list a base salary range by location and say the package also includes sign-on payments and RSUs, without amounts.

New York and San Francisco ranges sit above Seattle-area ranges for the same title. Applied Science intern postings show a single annualized starting pay instead of a range, so they are left out. Ranges seen October 29, 2026.

Amazon applied scientist interview process

Every step below comes from Amazon's Applied Scientist prep page and its virtual interview page, except where a stage notes candidate reports or the 2022 Amazon Science article.

Based on Amazon's own pages and candidate reports.

Amazon says to expect an outcome within 5 business days of the loop. The round count is where sources differ: Amazon lists four interviews, while a new-grad candidate described five virtual rounds (three fully behavioral, two with ML system design and coding), so ask your recruiter for your schedule. Last reviewed October 2026.

What is different at Amazon

• Amazon has separate prep pages for Applied Scientist and Research Scientist. The Applied Scientist loop lists problem solving and coding where the Research Scientist loop lists experimental design and data-driven modeling and scripting.
• Coding is weighted for a reason. Amazon Science says applied scientists are heavily involved in putting their algorithms into production, and the prep page links a sample coding question (recursive staircase).
• The Leadership Principles are scored in every interview, not in one behavioral round. Each interviewer typically asks two or three behavioral questions, and a Bar Raiser from outside the hiring team, one of more than 3,600 trained volunteers, decides with the hiring manager.
• Titles step up from Applied Scientist to Applied Scientist II, Applied Scientist III, Senior and Principal. Postings state no L levels. Data Scientist is a separate job category, and Machine Learning Engineer is an engineering title (for example on SageMaker) that is not an Applied Science role.
• Teams are spread across the company. US postings seen on October 29, 2026 cover Amazon Ads, Prime Video, AWS (including Neuron and the Center for Quantum Computing), Amazon Robotics, Stores and AGI and AI security.
• Amazon publishes no candidate AI rule on its careers pages. It says its new software development engineer assessment includes an interactive AI assistant, but ties that to SDEs, not applied scientists. A March 2025 news report described internal guidance asking candidates not to use generative AI tools during interviews unless explicitly permitted, which is a report and not a published Amazon rule.

Amazon applied scientist interview questions

Questions from our bank, then practice prompts we wrote for this guide around Amazon's work.

1from typing import Dict, List, Tuple
2
3def incrementality(rows: List[Tuple[str, str, int]]) -> Dict[str, dict]:
4    """Return incremental conversions and lift per campaign."""
5    pass

1from collections import defaultdict
2from typing import Dict, List, Tuple
3
4def incrementality(rows: List[Tuple[str, str, int]]) -> Dict[str, dict]:
5    users = defaultdict(lambda: {"treatment": 0, "control": 0})
6    conv = defaultdict(lambda: {"treatment": 0, "control": 0})
7    order = []
8    for campaign, group, converted in rows:
9        if group not in ("treatment", "control"):
10            raise ValueError(f"unknown group {group!r}")
11        if converted not in (0, 1):
12            raise ValueError(f"converted must be 0 or 1, got {converted!r}")
13        if campaign not in users:
14            order.append(campaign)
15        users[campaign][group] += 1
16        conv[campaign][group] += converted
17    out = {}
18    for campaign in order:
19        n, c = users[campaign], conv[campaign]
20        if n["treatment"] == 0 or n["control"] == 0:
21            continue
22        rt = c["treatment"] / n["treatment"]
23        rc = c["control"] / n["control"]
24        out[campaign] = {
25            "incremental": round(c["treatment"] - rc * n["treatment"], 1),
26            "lift": round((rt - rc) / rc, 3) if rc > 0 else None,
27        }
28    return out
29
30# Complexity: O(n) time for n rows, O(c) space for c campaigns.

How to prepare

A plan for the Amazon loop, most important first:

1. Write two stories for each Leadership Principle you expect, with metrics, since every interviewer asks two or three behavioral questions.
2. Review ML breadth and the depth of every model on your resume: losses, assumptions, failure modes and how you evaluated them.
3. Practice medium coding problems in Python with tests, starting with the recursive staircase sample on Amazon's prep page.
4. Prepare a tech talk on one past project or paper, covering the customer problem, the method, the result and what you would change.
5. Practice two ML system design problems end to end, such as fraud or spam detection, since candidate reports describe them in the loop.
6. Ask your recruiter how many loop interviews you have, whether a tech talk and a writing exercise apply, and what the AI rule is.

Practice path

Courses that cover the rounds above:

• Machine Learning
• ML System Design
• Data Structures and Algorithms
• Deep Learning
• Advanced Machine Learning
• Applied Statistics