How do I prepare for the SQL round of an Amazon interview in two weeks?

Work the 34 Amazon-tagged challenges in pattern order, about three a day: days 1-4 joins and aggregation (anti-joins, conditional aggregation, relational division), days 5-8 window functions (running totals, rolling averages, LAG, NTILE, DENSE_RANK), days 9-12 CTEs and ranking (multi-CTE pipelines, ROW_NUMBER top-N, gaps and islands, recursive CTEs), and days 13-14 a timed mock with no hints followed by a cold redo of anything you missed. Your first 10 challenge solves are free and can go on the 7 Medium challenges and the 3 Hard previews; Pro opens the rest of the Hard set. After each query, practice a one-sentence Dive Deep answer: what decision the number drives and what you would check before trusting it.

Does this page know what SQL questions Amazon asks?

No, and it does not claim to. Amazon does not publish its interview questions, and we have no dated, citable source for them — so this page states no round length, no tool and no process. What it has instead is the 34 challenges tagged Amazon in the SQL Quest bank (7 Medium, 27 Hard), chosen because their SQL matches the analytics work, with the reference solution shown once you solve one. The one page on this site that does carry dated public sources for a company's screen is the Capital One page.

Which SQL skills does the Amazon practice set cover?

Resolving every challenge tag onto the nine canonical skills the app's radar uses, the 34 tagged challenges lean on Querying Basics (34), Aggregation & Grouping (25), Subqueries & CTEs (21) and Window Functions (16) — counts, not shares of an interview. A challenge exercises several skills, so the counts add up to more than 34. The composition section on this page shows all 8 skills the set touches.

Are the Amazon practice challenges free?

Your first 10 challenge solves are free, with no signup — pick them from the 7 Medium challenges in this set or its 3 Hard previews. Pro ($29/mo or $99/yr) opens all 34, and the rest of the 300+ bank. Every challenge runs in the browser and shows its solution once solved.

What SQL dialect should I practice for Amazon?

We have no source for what Amazon runs internally, so this page does not guess at it. Practise standard SQL: everything here runs on SQLite in the browser, and joins, GROUP BY, CTEs, window functions and date logic transfer unchanged to PostgreSQL, MySQL, Snowflake and BigQuery. Vendor-specific syntax is rarely what an analytics question is actually testing.

Amazon SQL Interview Questions + 2-Week Prep Plan

How the interview runs (sourced)

The Amazon SQL interview, as candidates report it

Amazon does not publish the format. Each row is what candidates and prep guides have described publicly, with its source; where sources disagree, the row says so. Formats change — treat every specific as reported, not official. The sources below describe Business Intelligence Engineer / data analytics roles; other teams at Amazon may run a different loop.

Stages	A recruiter screen of about 30 minutes, one or two technical phone screens, then a virtual onsite loop of about five rounds.[1][2]
Round length	One candidate report puts the technical phone screens at 75 minutes each and the onsite rounds at roughly 60 minutes. A prep guide describes the onsite as five to six rounds of 45 minutes to an hour.[1][2]
Environment	The same report describes SQL written on a "CoderPad-style" editor where you often cannot execute the query, so correctness has to be reasoned rather than run.[1]
What the SQL round covers	Joins, window functions and ranking, with follow-ups on edge cases; the guide says to expect at least five questions across SQL, Python, visualisation and business analytics in the technical screen, without splitting out how many are SQL.[1][2]
Alongside the SQL	Leadership Principles come up throughout, including inside technical rounds — not only in the behavioural ones.[2][1]

[1] Blind — "Amazon Business Intelligence Engineer (BIE L5) Interview 2025" (candidate report, replies to 15 Oct 2025) (14 Sep 2025)
[2] Exponent — Amazon Business Intelligence Engineer interview guide (shown as "updated 2 months ago" on 14 Sep 2026)

Sources: Blind — "Amazon Business Intelligence Engineer (BIE L5) Interview 2025" (candidate report, replies to 15 Oct 2025) (14 Sep 2025); Exponent — Amazon Business Intelligence Engineer interview guide (shown as "updated 2 months ago" on 14 Sep 2026). Accessed 14 Sep 2026.

Preparation

SQL preparation for an Amazon interview — a 2-week plan (September 2026)

All 34 Amazon-tagged challenges, about three a day, in the order the patterns build on each other. Every solved challenge shows its solution. Cards marked Free are open to free accounts — your first 10 challenge solves are free, enough for the 7 Medium challenges and the 3 Hard previews — and Pro opens every card. Queries run in the browser — no setup, no signup to start.

Amazon publishes its Leadership Principles, and behavioral follow-ups in data loops are commonly framed around them. So after every query in this plan, practice a one-sentence Dive Deep answer: what decision this number would drive, and what you would check before trusting it. We do not know which questions any loop uses; these are the patterns that show up in analyst-round questions.

Days 1–4 · Joins & aggregation

Analyst-round questions start with counting the right thing: who never ordered, how orders split by status, who bought in every category. Get the join direction and the HAVING clause right before touching window functions.

MediumFree

Day 1 · Customers Who Never Ordered

Measures: Joins

MediumFree

Day 1 · Inactive Customers by Tier

Measures: Aggregation & Grouping · Joins · Subqueries & CTEs

MediumFree

Day 2 · Pivot: Order Status by Country

Measures: Aggregation & Grouping · Conditional Logic

MediumFree

Day 2 · Conditional Counting with CASE

Measures: Aggregation & Grouping · Conditional Logic