Amazon hires more software engineers in India than any other FAANG company — and it's also one of the most systematically preparable. Unlike Google (heavy competitive programming) or Microsoft (moderate DSA + open discussions), Amazon has a predictable structure: good DSA + strong Leadership Principle stories = offer. Every round follows the same template.
This guide covers everything: the exact 5-round process, DSA patterns Amazon loves, all 16 Leadership Principles with sample STAR stories, the Bar Raiser explained, system design for Amazon scale, and salary from SDE1 through Principal SDE in India.
1. Amazon India Interview Process — All 5 Rounds Explained
Amazon's interview process in India is called the "loop." For SDE1 and SDE2 roles, you'll typically go through an Online Assessment, a phone screen, and then 3–4 loop rounds in a single day (or over 2 days). Every loop round includes both technical and LP questions.
- 2 DSA coding problems: typically Medium difficulty; occasionally 1 Easy + 1 Hard
- Work style assessment: 15–20 LP-style situational questions (no right/wrong, but shows your LP alignment)
- Debugging section (some roles): 5–7 code snippets with bugs to identify
- MCQ section (some roles): OS, networking, or system concepts
- Most important: pass all test cases — partial credit is tracked but OA is generally pass/fail
- Top Amazon OA topics: sliding window, two-pointer, BFS/DFS on graphs, DP on arrays
- 1 DSA problem: Medium difficulty; you're expected to reach optimal solution
- 2 LP questions: Short answers; interviewer is looking for STAR structure and specific details
- Think aloud through your approach before coding — interviewers want to see your process
- Ask clarifying questions about constraints (input size, null handling, duplicates)
- Discuss time/space complexity after you write the solution
- Common LP for screen: Customer Obsession, Ownership
- 1–2 DSA problems: Medium to Hard; may be a follow-up extension of the first problem
- Extension pattern: solve first problem → interviewer says "now what if the input is a billion numbers?" → now test your optimisation thinking
- 2–3 LP questions: Dive Deep, Bias for Action, Deliver Results are common here
- For SDE2+: expect at least one question about how you handled conflict within a team
- Code cleanly — Amazon values readability; name variables well, add comments for tricky parts
- SDE2 and above: System design — typically a large-scale distributed system (see Section 3)
- SDE1: Second coding problem set; may include OOD (object-oriented design) questions
- For system design: start with requirements clarification, then API design, then data model, then architecture
- Amazon loves: exactly-once delivery guarantees, idempotency, graceful degradation under failure
- LP for this round: Invent and Simplify, Think Big, Earn Trust are common
- The Bar Raiser is an independent interviewer from a different team who has veto power over the hiring decision
- Their goal: ensure you raise the overall bar at Amazon — not just "good enough"
- Expect deeper probing: "In your story from round 2 about Customer Obsession, what would you have done differently?"
- May ask a DSA problem, but often focuses more on LP depth
- Be consistent: Bar Raisers compare notes with loop interviewers — inconsistencies are red flags
- Show long-term thinking: "I built this for scale from day one because…" > "I'd refactor it later"
- Contradicting yourself or being vague is the #1 reason for Bar Raiser rejections
2. All 16 Amazon Leadership Principles — With Interview Questions
Amazon has 16 Leadership Principles (LPs). Every round includes LP questions — usually 2–3 per round. You need at least 5–6 strong STAR stories that can be mapped to multiple LPs. Here's what each one means and the most common interview question for each:
Sample LP Story: Ownership + Deliver Results
Our team's payment processing service started failing 0.3% of transactions during peak load. The oncall engineer escalated to me even though payment infra was owned by a different team.
Identify root cause and fix it without waiting for the other team — every 0.1% failure was ₹2 Cr/month in lost revenue for our client.
Pulled CloudWatch logs myself, traced errors to a thread pool exhaustion under 5000 concurrent requests. Implemented an async processing queue with a circuit breaker pattern — deployed to staging, tested, got approval, and pushed to prod in 6 hours. Wrote an RCA document and shared it with both teams.
Transaction failures dropped to 0.01% within 1 hour of deployment. The circuit breaker pattern I built was adopted by 3 other services. My work directly saved ~₹5.8 Cr/month in transaction losses and the RCA became a team wiki article used in future oncalls.
Prepare 5–6 core stories, each richly detailed. Each story should flex across 2–3 LPs. For example, the story above can answer: Ownership, Deliver Results, Dive Deep, Bias for Action, and Customer Obsession. When an interviewer asks a new LP, pull the right story from your bank and adapt the framing. Practice until you can tell each story in 2 minutes and answer follow-up probes for 5 more minutes.
3. DSA Patterns Amazon India Tests Most Frequently
| Topic | Frequency | Key Patterns | Must-Solve LeetCode Problems |
|---|---|---|---|
| Arrays & Strings | Very High | Two-pointer, sliding window, prefix sum | 3Sum, Longest Substring Without Repeat, Container With Most Water, Product of Array Except Self |
| Trees & BST | Very High | BFS, DFS, level order, LCA | Binary Tree Max Path Sum, Lowest Common Ancestor, Serialize/Deserialize, Vertical Order Traversal |
| Graphs | High | BFS shortest path, DFS, Union-Find | Course Schedule, Number of Islands, Word Ladder, Clone Graph, Alien Dictionary |
| Dynamic Programming | High | 1D DP, 2D DP, interval DP | Coin Change, Word Break, Edit Distance, Partition Equal Subset Sum, Burst Balloons |
| Heaps / Priority Queue | Medium | Top-K, merge K sorted, streaming median | Find Median from Data Stream, Kth Largest Element, Top K Frequent Elements, Meeting Rooms II |
| Linked Lists | Medium | Fast/slow pointer, reversal, merge | Reverse Nodes in K-Group, LRU Cache, Merge K Sorted Lists, Flatten Multilevel List |
| Stack & Queue | Medium | Monotonic stack, design problems | Largest Rectangle in Histogram, Daily Temperatures, Trapping Rain Water, Min Stack |
| Backtracking | Lower | Subsets, permutations, combinations | Subsets II, Permutations, Word Search, N-Queens, Combination Sum |
Amazon's OA pulls heavily from a known set — filter LeetCode by "Amazon" tag and solve the top 100 by frequency. The loop rounds tend to use variants of the same problems. If you've seen 200+ Amazon-tagged LeetCode problems, you'll recognise the pattern instantly in the interview room.
4. System Design for Amazon India Interviews (SDE2+)
Amazon system design questions are grounded in their real products — AWS, marketplace, logistics, Prime Video, Alexa. The key Amazon-specific concerns are: scale (millions of events/second), reliability (five 9s availability), eventual consistency, and exactly-once processing.
Most Common Amazon System Design Questions in India
| Design Problem | Key Concepts to Cover | Amazon Angle |
|---|---|---|
| Design Amazon's Order Management System | State machine, idempotency, event sourcing, eventual consistency | Exactly-once order processing, failure handling, order status broadcast |
| Design a Key-Value Store (like DynamoDB) | Consistent hashing, replication, read/write quorum, compaction | Understand Dynamo paper — Amazon invented it |
| Design a Rate Limiter | Token bucket, sliding window counter, Redis distributed locks | API Gateway rate limiting at AWS scale |
| Design Amazon's Product Search | Inverted index, ranking (TF-IDF, BM25), Elasticsearch, caching | Personalization signals, seller-side ingestion pipeline |
| Design Amazon Prime Video Streaming | CDN, adaptive bitrate streaming (HLS/DASH), video transcoding | Regional CDN placement, content licensing geography |
| Design a Notification System | Fan-out, pub-sub (SNS/SQS), push vs pull, delivery guarantees | Amazon SNS is their product — show understanding of its design |
Amazon System Design Framework (Use This Structure)
- Requirements clarification (5 min): Functional requirements, non-functional (scale, latency, availability). Ask: "How many daily active users? What's the peak events/sec? Strong or eventual consistency?"
- Capacity estimation (3 min): QPS, storage, bandwidth. Show you can back-of-envelope estimate.
- API Design (5 min): Define the key REST or gRPC endpoints with request/response schemas.
- Data Model (5 min): SQL vs NoSQL choice with justification. Schema for core entities. Indexing strategy.
- High-Level Architecture (15 min): Draw the system: clients → load balancer → service layer → cache → database → message queue. Explain each component choice.
- Deep Dive (15 min): Pick 1–2 hard sub-problems and design them deeply. Amazon interviewers will guide you here.
- Trade-offs and Failure Handling (5 min): What breaks under load? What happens if the database goes down? How do you recover? This is where you show senior thinking.
5. Amazon India Salary Breakdown 2026 — SDE1 to Principal SDE
Amazon India compensation has three parts: base salary + signing bonus + RSUs (restricted stock units). The RSU vesting schedule is unusual: 5% in year 1, 15% in year 2, 40% in year 3, 40% in year 4. This makes year 1 and 2 total CTC lower than years 3 and 4.
Amazon's 5/15/40/40 vesting means your year-1 and year-2 total compensation is significantly lower than peers at Google or Microsoft (who use 25/25/25/25 or 33/33/33 vesting). Many Amazon India engineers leave after year 2 when vesting kicks in. If you join, plan to stay through year 3 — that's when your total comp spikes significantly. Factor this into salary negotiations and request a larger signing bonus to compensate for the front-loaded low vesting.
6. 8-Week Amazon India Prep Plan
- Revise arrays, strings, linked lists, stacks, queues — 30 LeetCode Easy/Medium
- Write your 5 core LP stories — one each for: Ownership, Customer Obsession, Deliver Results, Dive Deep, and Disagree & Commit
- Practice each story out loud — time yourself to 2 minutes per story
- Read "Working Backwards" (the Amazon book) to understand their culture deeply
- Binary trees, BSTs, BFS/DFS — 25 Medium problems from Amazon tag on LeetCode
- Graph problems: course schedule, number of islands, word ladder, alien dictionary
- DP: coin change, word break, edit distance, partition equal subset — 20 problems
- Expand LP story bank to cover remaining LPs: Invent & Simplify, Bias for Action, Frugality
- Design: Order management system, key-value store, rate limiter, notification system
- Study the original Amazon Dynamo paper (summary version on Medium)
- Mock LP interview: ask a friend to play interviewer and drill you on follow-up questions 4 levels deep
- Solve 20 more LeetCode Hard problems from Amazon tag
- Practice system design out loud with a 45-minute timer
- Take 2–3 Pramp mock interviews (coding + behavioral) — Amazon-focused if available
- Simulate the OA: 2 LeetCode Mediums under 90 minutes, no hints
- Revise all LP stories — ensure consistency across all 16 LPs
- Apply through Amazon.jobs (direct) + referral if available
- Research your specific target team: AWS, Alexa, Amazon Pay India, Flipkart (Amazon acquired Cloudtail) — teams differ in culture and design focus
- Prepare salary research: check Levels.fyi for Amazon India SDE1/2 data points
7. Amazon-Specific Interview Do's and Don'ts
✅ What Amazon Interviewers Love
- Data-driven answers: "Latency improved by 35%" > "performance got better." Always quantify.
- Taking ownership in LP stories: Use "I" not "we." What did YOU specifically do?
- Thinking about the customer: Even in DSA problems, ask about edge cases that affect user experience.
- Clean, readable code: Name variables like
maxProductAtIndexnotmax2. Add a comment if logic isn't obvious. - Admitting failure with learning: "I made a mistake, here's what happened, here's what I changed" is a Bar Raiser hire signal.
❌ What Amazon Interviewers Hate
- Vague LP answers: "my team fixed the issue" — Who did what? What were the specifics?
- Jumping to code without clarifying requirements
- Using jargon without explaining it — "we used microservices" without explaining the trade-off decision
- Inconsistency between LP stories across rounds — Bar Raisers will catch this
- Never asking for feedback during the interview — show intellectual curiosity
- Blaming others in LP stories — even subtly
8. Frequently Asked Questions
⚡ Crack the Amazon OA with PrepFlix DSA Tracks
Practice structured DSA problems organized by Amazon's most-tested patterns. Arrays, Trees, Graphs, DP — everything you need to pass the Online Assessment and Loop rounds.
Start Practicing Now →