Amazon India Interview Questions 2026: Leadership Principles

Amazon India SDE interview process (SDE-1/SDE-2/SDE-3):

1. Online Assessment (OA): 2 coding problems (90 minutes, LeetCode medium-hard difficulty), sometimes a third problem or a work simulation (scenario-based questions on debugging or output prediction). Automated scoring; top candidates advance.

1. Phone Screen (for experienced hires): 45-60 minutes; 1-2 coding problems plus 1-2 Leadership Principle questions via a shared coding environment.

3. Onsite Loop (3-5 rounds, each 45-60 minutes): - Round 1: Coding (2 LeetCode medium problems; arrays, trees, DP) - Round 2: Coding plus LP (1 coding problem plus 2-3 LP questions) - Round 3: System Design (SDE-2 and above) - Round 4: LP-focused (2-3 in-depth LP deep-dives) - Round 5 (sometimes): Bar Raiser round (an Amazon employee from another team who votes independently)

1. Hiring Decision: Bar Raiser's vote is counted equally with loop interviewers. A negative Bar Raiser vote typically means a reject regardless of other votes.

Amazon has 16 Leadership Principles (LPs). Most frequently asked:

1. Customer Obsession: 'Tell me about a time you went above and beyond for a customer.' The customer can be internal (another team relying on your API) or external. Focus on: what the customer's actual need was (not what they asked for), what you specifically did that was not required, and the measurable outcome.

1. Ownership: 'Tell me about a time you took ownership of something outside your responsibility.' Show: you spotted a problem no one else was addressing, why you decided to own it, what you did, and what the outcome was. Amazon values engineers who do not say 'that is not my job.'

1. Invent and Simplify: 'Tell me about a time you simplified a complex process or system.' Show efficiency improvement, complexity reduction, or an elegant solution. Quantify: 'reduced deployment time from 2 hours to 8 minutes', 'eliminated 3 manual steps in the release process.'

1. Dive Deep: 'Tell me about a time you had to investigate a problem deeply to find the root cause.' Show analytical rigor: you did not accept the first explanation; you dug into logs, metrics, and code until you found the real cause.

1. Bias for Action: 'Tell me about a time you made a decision with incomplete information.' Show you can make calculated decisions under uncertainty without waiting for perfect data. Include the risk you identified and how you mitigated it.

Preparationwrite 5-7 STAR stories and map each to 2-3 LPs. Stories should be specific, quantified, and YOUR contribution (not 'we').

Amazon India most common coding topics:

Arrays'Two Sum (HashMap)', 'Best Time to Buy and Sell Stock (one-pass min tracking)', 'Merge Intervals (sort by start, merge overlapping)', 'Trapping Rain Water (two pointer or stack)'.

Trees'Binary Tree Level Order Traversal (BFS + queue)', 'Validate Binary Search Tree (in-order traversal or range check)', 'Lowest Common Ancestor (recurse; if both found in different subtrees, current node is LCA)', 'Serialize and Deserialize a Binary Tree'.

Strings'Longest Palindromic Substring (expand around center)', 'Longest Substring Without Repeating Characters (sliding window)', 'Valid Anagram (character frequency map)'.

Dynamic Programming'Coin Change (minimum coins: BFS or bottom-up DP)', 'Longest Increasing Subsequence', 'Word Break (DP with memo)'.

Amazon-specific patterns: graph problems appear more at SDE-2+ (course schedule, word ladder). Amazon tends to ask slightly more DP than Google. The OA has one harder and one easier problem; solve the easier one completely before spending time on the harder one.

Amazon India system design questions:

1. 'Design Amazon's order management system': State machine: placed, payment_pending, payment_confirmed, processing, shipped, out_for_delivery, delivered, cancelled, returned. Each state transition publishes an event to Kafka. Downstream services (inventory, seller, logistics, notification) consume the event independently. Idempotency: each state transition checks the expected previous state before applying (prevents double-processing if the event is consumed twice).

• 2. 'Design a URL shortener at Amazon scale (100M URLs, 1B redirects per day)': API: POST /shorten (takes long URL, returns short code), GET /{code} (redirects to long URL).
• Storage. base62 encoding of auto-increment ID (no collisions, shorter than hash).
• Caching. Redis for hot short codes (80% cache hit on the top 5% of URLs).
• Database. read replicas for the redirect path.
• Analytics. async click tracking via SQS consumer (do not slow down the redirect).
• Response code. 302 (not 301) so the redirect is not cached forever by browsers (allows URL expiry or update).

• 3. 'Design Amazon's notification service (email/push/SMS) at scale (100M+ notifications per day)': Fanout pattern: a single order event may trigger 3 notifications (email, push, SMS). Kafka partitioning by user_id ensures per-user ordering.
• Worker pools. separate consumer groups per channel (email worker, push worker, SMS worker) for independent scaling.
• Delivery. exponential backoff for failed deliveries; dead letter queue (DLQ) for notifications that fail after N retries.
• Deduplication. idempotency key per notification to prevent duplicates on retry.

The Amazon Bar Raiser round:

What the Bar Raiser evaluates: they are looking for candidates who will raise the average quality of Amazon's workforce. They ask LPs at a deeper level than the standard loop. They may probe for inconsistencies in your stories or ask follow-up questions that reveal whether your answers are genuine experiences or practiced scripts.

Common Bar Raiser techniques: - Asking you to describe the same situation from a different perspective: 'You said your manager disagreed with your approach. What did they see that you initially missed?' - Going deeper on specifics: 'You mentioned you improved latency by 40%. How did you measure that? What was the baseline? What exactly did you change?' - Testing LP consistency: 'Earlier you said you prioritise customer obsession. In this situation, you chose not to involve the customer. Help me understand that decision.'

How to prepareyour STAR stories must be genuinely your experience, deeply remembered, and quantified. Rehearsed-sounding stories are obvious. The Bar Raiser is experienced at spotting them.