Backend interviews are the closest to a "classic SDE" loop, but they have their own centre of gravity: solid DSA plus a real expectation of system design, databases, and concurrency. Grind only LeetCode and you'll clear the coding round but stall on the design conversation — which is often where backend offers are won or lost.

Frontend goes wide-and-shallow on DSA; backend goes a notch deeper on DSA and adds an entire second axis — designing systems that don't fall over.

The DSA that recurs in backend loops

• Hash maps & sets — caching, dedup, lookups; everywhere in backend work and interviews.
• Trees & graphs — dependency resolution, routing, hierarchies; BFS/DFS, topological sort show up more than in frontend.
• Heaps / priority queues — schedulers, rate limiters, "top-K" problems.
• Two pointers, sliding window, binary search — still high-frequency.
• Medium DP & greedy — occasional; learn after the above is solid.

The other half: design & data

This is what separates backend prep from generic DSA grinding. Budget real time for:

• Database & SQL — indexing, joins, transactions, N+1 queries, when to denormalise.
• System design basics — caching, load balancing, queues, rate limiting, idempotency, how to scale a read-heavy vs write-heavy service.
• Concurrency — race conditions, locks, why your in-memory counter breaks across workers.
• APIs — REST design, status codes, pagination, auth.

A realistic backend split

Area	Weight	Why
Core DSA	45%	Clears the coding round
System design + databases	40%	Where backend offers are decided
Concurrency + APIs	15%	The "have you actually built backends" signal

Pick Backend in SUITS and the curated 370 problems reorder to the patterns backend loops over-weight — so you're not grinding segment trees before hash maps and graphs. For the cross-role view, see the DSA roadmap by role, and compare with the frontend DSA breakdown if you're still choosing a track.