SDE-1 All-In Guide
System Design: Track F

System design: the HLD track

What high-level design interviews ask, how they differ from LLD, and the weekly order to study them.

This track covers high-level design (HLD), the round where you sketch a whole system: servers, databases, caches and queues. Read this page once now, then use it to place each video and each case study in a week. The low-level design track covers the classes and code, and the two are graded differently.

Priority for SDE-1

DSA and your projects come first. HLD gets 2 to 3 hours a week until week 20, and it only grows after that. A finished Ledgerline answer beats a memorised URL shortener.

What HLD actually is

An HLD answer describes how a system is put together at the level of machines and services. You choose a database, add a cache, split traffic across servers, and explain what happens when one part dies. You do not write classes or methods.

LLD answers a different question. It asks how the code inside one service is organised, and it is usually judged on whether your code runs. The table below is the whole difference in four rows.

QuestionLLDHLD
Unit of designClasses and interfacesServices, stores, queues
OutputRunning code, 90 minutesBoxes and arrows, 45 minutes
Judged onExtensibility and testsTradeoffs and failure modes
Weak answer looks likeClean code that does not runA diagram with no numbers

What an SDE-1 loop asks in India

Be honest about this: most SDE-1 loops do not run a full 45-minute HLD round. The common patterns are a 20-minute discussion of your own project, a few HLD questions folded into the LLD round, or one round that starts with "design a rate limiter" and drifts into schema and caching.

That is good news for scheduling. You do not need a course on distributed consensus in week 4. You need enough vocabulary to answer project questions without hand-waving, and one rehearsed design you can draw from memory.

SDE-2 loops go deeper. The interviewer pushes on capacity numbers, replication lag, hot partitions and what breaks at 10x traffic. The same page prepares you for that, one level at a time.

What SDE-2 expects

At SDE-2 the interviewer already assumes you can draw the boxes. They test whether your choices survive load and failure:

  • You size the system out loud: QPS, storage, and the read to write ratio.
  • You name the bottleneck before they point at it.
  • You say what happens when the cache dies or the primary fails.
  • You defend one tradeoff and admit what it costs.

None of this needs a job title. It needs practice at narrating decisions, which the case studies page drills.

The level legend

Every page in this section tags each item with one of three labels. Use them to decide what to skip on a tired night.

LabelMeaningWhat you do with it
SDE-1 mustNeeded for the loops this plan targetsLearn it before week 20
SDE-1 niceUseful if time allowsRead after the must list
SDE-2 laterDepth for after an offerSkim now, return later

If a page mixes levels and you cannot tell what to skip, that page needs a retag. Tell the person maintaining it.

How to study a video

Watching is the weakest form of study. A video you only watched is gone in a week. Use this loop instead, and it takes about the same time as watching twice.

Watch at 1.25x. Go to 1.5x for anything you already know, such as HTTP basics.

Pause at every diagram. Count the boxes and say what each arrow carries.

Redraw the diagram from memory on paper, then compare. Fix only what you got wrong.

Write three lines of notes: the problem, the one idea, and where you would use it.

Retrieve it a week later. Close the notes and redraw the diagram again. If it comes back, it is yours.

The three-line note is the unit of revision. A page of notes you never reopen is a page you wasted.

The weekly cadence

Keep system design to a small weekly block. Add practice as interview needs become clear.

WeeksWhat you doHours a week
1 to 3Focus on DSA, the first project, and fundamentals0
4 to 25Building blocks, selected videos, and a case study when readyUp to 2
26 to 34Review around interview loopsWithin the available study block

Building blocks start in week 4, when Ledgerline's transactions give you a concrete example. Choose videos that support the current topic; do not treat the full playlist as a required course.

Case studies can start when you can explain the basic building blocks. Use one case study at a time and keep this track inside its weekly time limit.

Loading diagram…

That is the whole order. Skipping to case studies feels productive and produces answers with no numbers in them.

The pages in this section

The System Design Primer is the free written backup for the whole track. Use it when a video moves too fast, not as a third course.

Two habits worth building now

The first is tying every answer to a project. "In Ledgerline I used an outbox so a crash could not lose a transfer" is stronger than any definition you can recite.

The second is saying the tradeoff out loud. Every design choice buys something and costs something. An answer that only lists benefits sounds like a tutorial, and interviewers hear that immediately.

Where the written depth lives

Hello Interview's system design guide shows how strong candidates structure an answer. Read one page of it after your first mock, when you have your own attempt to compare.

Common mistakes

  • Start with requirements and numbers, even if the interviewer seems in a hurry.
  • Draw the simple version first, then add a cache or queue only when a number demands it.
  • Use your own projects as examples. They are the only systems you can answer questions about.
  • Rehearse the same design until it takes 30 minutes with no notes.

On this page