1600+ Real Interview Questions from 50+ Top Companies

Q: Design a URL shortener that handles 10 billion clicks per month

Requirements: Read-heavy (100:1 ratio), low latency (<10ms), high availability (99.99%). Encoding: Base62 gives 3.5 trillion unique combinations. Architecture: API Gateway, Write Service, Cassandra for writes, Redis cache for hot reads, CDN for geographic distribution, consistent hashing for shard routing.

Q: Implement an LRU cache with O(1) get and put operations

Use a HashMap for O(1) lookup combined with a Doubly Linked List for O(1) insertion/deletion. HashMap stores key to node pointer. Get: move accessed node to head. Put: add new node to head, evict tail if at capacity. Time: O(1) for both operations. Space: O(capacity).

Q: Describe a time you simplified a complex system

Amazon Leadership Principle: Invent and Simplify. S: Inherited a payment reconciliation system with 12 microservices and 47 configuration files. T: Reduce operational burden — team spent 30% of time on maintenance. A: Consolidated into 3 services with shared config layer, automated deployment. R: 60% reduction in incidents, team velocity increased 2x.

Q: How would you design a real-time notification system for millions of users?

Requirements: <100ms delivery, millions of concurrent connections, multi-channel (push, email, SMS, in-app). Architecture: Event producers → Message queue (Kafka) → Notification service → Channel routers. Use WebSockets for real-time in-app, APNs/FCM for mobile push. Fanout pattern for broadcasts, per-user queues for targeted notifications.

Q: Given an array of integers, find two numbers that add up to a target sum

Use a hash map to store complements. For each number, check if (target - number) exists in the map. One pass solution. Time complexity: O(n). Space complexity: O(n). Python: def two_sum(nums, target): seen = {}; for i, num in enumerate(nums): if target-num in seen: return [seen[target-num], i]; seen[num] = i

Question 1

Behavioral Amazon SDE • L5 Medium

Tell me about a time you had to make a difficult technical decision under pressure.

STAR Format ~10 min Asked 47 times

Answer

S: Led migration of payment service during Black Friday freeze. Team of 6, $4M daily transaction volume at stake.

T: Needed to reduce P99 latency by 40% without any downtime or data loss — with only 48 hours before code freeze.

A: Proposed shadow deployment with 5% canary traffic. Got buy-in from team in 30-min design review. Personally monitored rollout with custom dashboards.

R: Achieved 52% latency reduction with zero incidents. Saved estimated $2M in potential Black Friday revenue loss.

🔒 Sign up free to see the full answer

Sign Up Free — 25 answers/month

Question 2

System Design Google SDE • L5 Hard

Design a URL shortener that handles 10 billion clicks per month.

Answer

Requirements: Read-heavy (100:1 ratio), low latency (<10ms), high availability (99.99%), 100M new URLs/month.

Encoding: Base62 (a-z, A-Z, 0-9) gives 62^7 = 3.5 trillion unique combinations — more than enough.

High-level architecture: API Gateway → Write Service → Database (Cassandra for writes) + Cache (Redis) for hot reads. CDN for geographic distribution. Consistent hashing for shard routing...

🔒 Full architecture breakdown inside

Sign Up Free

Question 3

DSA / Coding Meta SDE • L4 Medium

Given an array of integers, find two numbers that add up to a target sum.

Answer

Approach: Use a hash map to store complements. For each number, check if (target - number) exists in the map. One pass solution.

def two_sum(nums, target):

seen = {}

for i, num in enumerate(nums):

complement = target - num

if complement in seen: return [seen[complement], i]

seen[num] = i

🔒 Full solution + complexity analysis

Sign Up Free