Amazon filters on OA first, then loops DSA with Leadership Principles in every round. Bar Raiser is real. You need clean code under time and STAR stories with numbers - not buzzwords.
What actually rejects people
Failing OA test cases, or LP stories that are vague and unmeasurable.
First round: Amazon OA (2 DSA + work simulation + LP MCQs)
Time split that matches how candidates report outcomes (not a marketing pie chart):
DSA 55%
System design / CS 20%
Behavioral 15%
Mocks 10%
Patterns are from student reports and vary by role and drive. Trust your college placement email over blog folklore.
Stage
Time
What they check
Online assessment
70-120 min
2 coding + simulation/debug + LP MCQs
Technical loops
45-60 min each
DSA + LP probes in the same round
System design / depth
varies by level
SDE-1 light; SDE-2+ heavier
Bar Raiser
45-60 min
Independent bar check - consistency matters
HR / offer
20-30 min
Role, location, CTC
Master DSA (graphs, sliding window, bit manipulation, DP) - practice 150+ problems on LeetCode (Amazon tag)
Study system design basics (SDE-1) and scalable systems (SDE-2+)
Prepare STAR stories for all 16 Amazon Leadership Principles
Practice coding problems in Java/C++/Python
One timed block that matches Amazon’s first round (OA coding)
Log misses: topic + why you failed (speed, concept, carelessness)
Re-solve yesterday’s misses cold
Twice a week: 10 minutes of resume/project narration out loud
Week
Focus
Exit criteria
1-2
Arrays, hashing, two pointers, sliding window
40+ clean mediums
3-4
Trees, graphs, BFS/DFS, heaps
Implement BFS/DFS cold
5-6
DP + timed 2-problem mocks
Finish both with tests
7
System design lite + CS revision
Explain a URL shortener / newsfeed at SDE-1 depth
8
Behavioral + weak-topic closure
STAR bank ready; re-solve misses
Order problems by confidence after a 3-4 minute skim
Passing test cases beats a brilliant half-solution
Invent edge cases before submit; watch timeboxes
Restate → brute force → optimize → code → test
Speak complexity and trade-offs out loud.
Tie CS answers to a project when it strengthens the point
Why Amazon, why this team/role, location flexibility
STAR with a metric; pick 2-3 real company values and prepare proof
LeetCode Amazon-tag mediums: graphs, sliding window, heaps, DP
Timed 2-problem OA mocks until both pass with edge cases
STAR bank mapped to Customer Obsession, Ownership, Dive Deep, Deliver Results
Work-simulation style judgment: prioritize, write clear responses
One language you can debug live (Java/C++/Python)
Arrays, strings, hashing, trees, graphs
Complexity analysis without freezing
Two project explanations (2 minutes each)
DP patterns (knapsack, LIS, grid, string)
Heaps, tries, union-find as needed
Light system design for SDE-1
Full-length mock every other day
Light revision only - no new rabbit holes
Sleep, ID proofs, machine/network checklist for test day
30 days out: weekly full OA-style mocks; close weak DSA patterns; draft STAR stories.
7 days out: alternate mock / review days; re-solve misses cold; no new patterns.
Day before: skim notes, check machine/network/ID logistics, sleep.
Failing OA test cases, or LP stories that are vague and unmeasurable - fix this first
Preparing only interviews while failing Amazon’s first round
Collecting notes without timed mocks
Listing five frameworks on the resume and defending none
Changing language/stack one week before the drive
Silent coding - interviewers cannot grade your head
Memorizing solutions you cannot adapt
Google · Microsoft · Meta · Apple · Netflix · Flipkart
Next step: Run one timed mock this week that matches Amazon’s first round (OA coding), then build the weeks around your weakest section - not around what is most fun to study.