System design + LLD practice
System design scenarios show what broke in production. LLD challenges hand you real-looking code and ask you to spot inefficiencies and explain a better approach — with feedback in minutes.
~20 min · Learn: N+1 queries · fetch in a loop
Fix the broken code
Interview-style code reviews for 0–2 years — spot inefficient patterns and explain your fix.
Queries inside a loop
Learn: fetch in a loop
In QA, the dashboard feels slow when about fifty orders are on the screen. · Database activity spikes on each page load in the QA logs.One method doing too many jobs
Learn: split responsibilities
addToCart() has grown past four hundred lines and mixes several jobs in one place. · A pricing-rule change and a persistence change both land in the same method.Server must check client input
Learn: trust boundaries
QA saved orders with quantity -1 and an email that was not a real email address. · Support finds nonsense rows in the orders table.Database object returned as JSON
Learn: API boundary
A security review found a password hash field on a nested customer object in the JSON response. · The response payload grew large when an order had many line items.Same request charged twice
Learn: safe retries
QA replayed the same Pay request twice and saw two charges for one order. · A mobile retry after a slow network can create a second charge for the same order.Loading every row into memory
Learn: bounded reads
The search feels fine on a small development database with a few hundred users. · In QA with about ten thousand users, the heap spikes and the UI freezes.Fix the architecture
Real production stories — study flawed designs and explain how you would rebuild them.
Big Billion Days traffic overloaded Flipkart’s notification path, pushing the database to full CPU, delaying alerts up to 28 minutes, and sending some shoppers the same message more than once.
Learn: what breaks first
When Big Billion Days traffic hit, messages arrived 20–30 minutes late. · Some shoppers got the same notification two or three times.
Related practice problems
During Amazon India’s Great Indian Festival, an accidental cache clear sent almost every request to the database, filling connection pools, maxing out database CPU, and slowing the slowest requests to 30 seconds.
Learn: immediate triage
After the cache was cleared, about 95% of requests missed the cache for roughly 8 minutes. · API servers ran out of database connections — 480 of 500 slots were in use.
Related practice problems
During a Diwali sale week on Razorpay checkout, shoppers who saw Payment failed and tapped Retry were charged twice for the same order — about 2% of checkouts — and support tickets jumped about 4×.
Learn: double charge
The payment provider timed out with no clear success or failure on the first attempt. · Checkout showed Payment failed and offered Retry; shoppers who retried were often charged twice.
Related practice problems
On Razorpay’s shared API gateways, one misconfigured enterprise key used about 80% of capacity for 45 minutes, starving other merchants with about 22% too-many-requests errors.
Learn: fairness
For 45 minutes one API key used about 80% of shared gateway capacity. · Other merchants were starved — about 22% of their requests hit too-many-requests errors.
Related practice problems
On Instagram, a celebrity post to about 40 million followers overloaded one timeline shard at 100% CPU, slowing some feeds to about 45 seconds and leaving about 2.1 million fan-out writes waiting.
Learn: skew
A celebrity with about 40 million followers posted during peak hours. · One timeline shard hit 100% CPU while others stayed under 30%.
Related practice problems
On WhatsApp, a mobile network blip triggered a reconnect storm, duplicating about 12% of messages and pushing WebSocket gateway CPU to about 92%.
Learn: reconnect
A mobile network blip disconnected millions of users at once. · On reconnect, clients retried every message that had not been acknowledged yet.
Related practice problems
During a cricket World Cup campaign on bit.ly, one viral short link drove about 50× traffic, causing about 95% cache misses, redirect waits up to about 8 seconds, and saturated database replicas.
Learn: cache miss storm
A viral marketing link drove about 50× normal traffic to a single short code. · Cache miss rate for that code hit about 95%.
Related practice problems
During Swiggy’s dinner rush, a slow inventory service blocked shared gateway threads, returning bad-gateway errors on about 48% of requests for about 22 minutes while health checks still looked fine.
Learn: cascade
The inventory service slowed down — responses took up to about 12 seconds at worst. · The gateway returned bad-gateway errors for about 48% of requests over about 22 minutes.
Related practice problems
During Flipkart Big Billion Days prep, a deploy slowed workers to about 40% speed while producers kept publishing — consumer lag grew unnoticed for about six hours and the oldest jobs waited over four hours.
Learn: silent lag
After a deploy, workers slowed to about 40% of normal throughput. · Producers kept publishing at full speed.
Related practice problems
On Nykaa’s beauty-sale launch day, server HTML and browser JavaScript disagreed on cart state — hydration errors hit about 18% of sessions, and about 8% of users saw a white screen.
Learn: SSR/client split
Launch traffic drove about 4× visits to the new product pages. · Hydration errors hit about 18% of sessions within 20 minutes.
During Myntra’s End of Reason Sale, a 4.2MB JavaScript checkout bundle pushed time-to-interactive to about 14 seconds on mid-tier phones and raised abandonment about 31% while the payment API stayed healthy.
Learn: main-thread blocking
Checkout traffic rose from about 800 to about 2,400 requests per minute. · The main JavaScript bundle grew to about 4.2MB transferred.
During a Dream11 IPL ad blitz, a landing page without reserved space for hero, fonts, and ads hit a layout shift score of about 0.48, jumped about 400 pixels, and cut conversion about 22%.
Learn: visual stability
Paid traffic jumped from about 400 to about 6,200 requests per minute. · Layout shift score hit about 0.48 (good is under 0.1).
Complete the arc
Each pairing starts with a real production scenario, then continues with a greenfield system design prompt.
9 scenario + design pairings
Design a new system
Interview-style greenfield prompts — sketch requirements, APIs, and architecture, then get AI feedback. Browse featured starters here, or open the full problem library for filters and progress tracking.
Loading problems…