The HLD interviewThe round
Two architects get the same brief. Why does only one get hired?
0%
System design, one shard at a time
Tap a question.
Get it in 60 seconds.
Then go as deep as you like: a quick check, interactive deep dives and the full breakdown.
One from every module
Start with a question
- Core conceptsRetrograde
Why can adding servers make your system slower?
0% - Networking and APIsTTLs
You edited the DNS record. Why do users still reach the old server?
0% - Data and storageData models
Why does a good hospital file the same patients five different ways?
0% - Caching fundamentalsThe idea
Why does a tiny counter serve most of a kirana shop’s sales?
0% - Messaging and streamsLoad levelling
Why does a flash sale fill your queue but never time out checkout?
0% - Distributed coordinationSlow or dead
A heartbeat is 4 seconds late. Is the node dead, or just slow?
0% - Running in productionRetries
The payment app spins on Diwali evening. Do you press Pay again?
0% - Why is it slow? Performance anti-patternsThe symptom
Database CPU at 100%, app servers idle. Why don’t more pods help?
0% - SecurityRevocation
You hit log out. Why does a stolen copy of your token still work?
0% - Multi-tenant SaaSTenant vs user
One login opens 40 companies’ books. How do they never mix?
0% - Architecture stylesOne truth
Two people tap “Book” on the last seat. Who decides who gets it?
0% - System design trade-offsUp or out
Your dosa stall is swamped every morning. Bigger tawa, or more counters?
0% - Key technologiesOne thread
How does Redis serve thousands of clients on one thread, with no locks?
0% - Common patternsThe playbook
Sale day, 100,000 product views a second. Where do you even start?
0% - InfrastructureCounting
How does one API key sneak 10,000 requests past a limit of 100?
0% - Search & discoveryPrefixes
How does a search box guess “iphone 16” before your fourth letter?
0% - Storage & mediaShort codes
How do just seven characters name 110 billion links without a clash?
0% - Social & feedsCelebrities
A film star posts to 100 million followers. Why doesn’t your feed freeze?
0% - Real-time & collaborationPost office
Why does WhatsApp delete your message the moment it’s delivered?
0% - Location & mapsMatching
Why doesn’t Uber send you the closest driver?
0% - E-commerce & bookingFan-out
How does a product page ask six services and still answer in 200 ms?
0% - Payments & fintechUnknown outcomes
The bank debited you, but the app says “Payment failed”. Why?
0% - Data, analytics & monitoringExactly once
Ten billion ad clicks a day. How do you bill each exactly once?
0% - Developer platformsSandboxing
Every Submit runs a stranger’s code on your machines. What stops a fork bomb?
0% - AI & LLM systemsStreaming
Your phone drops from Wi-Fi to 4G mid-answer. Why isn’t a word lost?
0% - 980 more
Still curious?
Keep scrolling through questions from every topic, a page at a time.
More questions
All modules
26 modules · 252 topics
Every topic has a 60-second story, a quick check, interactive deep dives and the full written breakdown.
02Core concepts
Scale and performance
Reliability
Spreading data across machines
03Networking and APIs
How a request travels
API styles
Protecting the API
04Data and storage
Choosing a database
Schema design
Transactions and concurrency
Query performance
Scaling the data layer
Data structures at scale
Deep dives
06Messaging and streams
Queues and topics
07Distributed coordination
Agreeing and finding each other
Transactions and ordering
08Running in production
Resilience and recovery
Observability
Scaling the fleet
09Why is it slow? Performance anti-patterns
In the service
Under shared load
10Security
12Architecture styles
14Key technologies
20Real-time & collaboration
23Payments & fintech