01OverviewTypes of system design interviews

01 · Orientation

Types of system design interviews

Distinguish product, infrastructure, object-oriented, and ML system design prompts.
7 minConcept guideReference-informed · independently authored
01

Lesson spine

What you need to understand.

The same opening prompt can test very different skills; identify the interview contract before choosing your depth.

01

Product design

Start with actors, critical user journeys, core entities, public APIs, and scale. Examples include feeds, marketplaces, maps, and chat.

02

Infrastructure design

Define the primitive and its guarantees before products or vendors. Expect depth on partitioning, replication, concurrency, and recovery.

03

Object-oriented design

Model responsibilities, interfaces, composition, state transitions, and extension points. Distribution is usually secondary unless requested.

04

ML system design

Connect data collection, training, evaluation, deployment, online inference, monitoring, and feedback. Model quality is only one system metric.

05

Detect the format

Ask what level, traffic envelope, and primary use cases the interviewer wants. Their answer determines whether you should draw services, classes, or an ML lifecycle.

02

Before the boxes

Frame the decision.

Outcome

What must work

Distinguish product, infrastructure, object-oriented, and ML system design prompts.

Scale

What changes the design

Identify and confirm the format in the first 90 seconds.

Boundary

What earns a decision

Separate facts you learned from assumptions you made, then name the requirement that justifies each architectural choice.

Non-goal

What stays out of scope

Do not solve every adjacent product problem. State exclusions so the interview can go deep on the decisions that matter.

03

Architecture map

Trace the reasoning, not just the boxes.

Types of system design interviews · concept mechanism

Walk the method in order. At each arrow, state what new evidence or decision lets you advance without hiding an assumption.

  1. 01

    Prompt clues — Identify the expected artifact Make the output of this reasoning step explicit before moving on.

  2. 02

    Interview type — Product, infrastructure, OOD, or ML Make the output of this reasoning step explicit before moving on.

  3. 03

    Evaluation lens — Select the relevant signals Make the output of this reasoning step explicit before moving on.

  4. 04

    Method — Adapt the sequence Make the output of this reasoning step explicit before moving on.

  5. 05

    Checkpoint — Confirm the format Summarize the decision and invite the next deep dive.

04

Decision table

Make the trade-offs explicit.

DecisionDefensible positionCost to acknowledge
Primary methodProduct design rewards end-to-end flows; infrastructure design rewards invariants and recovery.A visible structure costs a little time but prevents a shallow or directionless design.
Breadth vs. depthCover the end-to-end shape, then spend the remaining time on the highest-risk requirement.Going deep too early can leave the core flow incomplete; staying broad hides engineering judgment.
Assume vs. clarifyAsk about decisions that change the architecture and state reasonable defaults for everything else.Too many questions consume the interview; silent assumptions make the solution impossible to evaluate.
05

Failure review

Recover the conversation.

NoticeClarifyReframeDecideSummarize

Topic-specific risk

A memorized template can answer the wrong kind of question.

Response

Return to the prompt, make the missing assumption explicit, and restate the decision in a form the interviewer can challenge.

Premature solutioning

Jumping to technology before requirements makes every later choice look arbitrary.

Response

Pause, restate the user flow and scale envelope, then connect each component to a named constraint.

Unstructured deep dive

A good idea can still read as weak judgment when the interviewer cannot follow the decision path.

Response

State the claim, alternatives, deciding constraint, and accepted cost before drawing the next box.

06

Evidence + level bar

Make the reasoning observable.

Core signals

Clarity of the argument

Make requirements, assumptions, alternatives, and accepted costs audible. The interviewer should never need to guess why a box exists.

Mid-level

Complete and clear

Finish the happy path, identify the state owner, choose reasonable building blocks, and explain one scale mechanism.

Senior

Trade-offs and failure

Separate read and write paths, define consistency, explain partitioning, and make duplicate or partial failure safe.

Staff+

Evolution and operations

Discuss multi-region boundaries, migration, tenant isolation, capacity, observability, and how the architecture changes over time.

07

Interview language

Open the deep dive with a claim.

“For Types of system design interviews, the method I want to make explicit is this: Product design rewards end-to-end flows; infrastructure design rewards invariants and recovery. I’ll state the output of each step, surface the assumption behind it, and use the highest-risk requirement to choose where we go deep.”

08 · Retrieval check

Can you defend it without the page?

  1. For Types of system design interviews, where is the correctness boundary and which failure would you test first?
  2. Which assumption most changes the design, and how would you validate it with the interviewer?
  3. Where would you stop covering breadth and begin the highest-value deep dive?
  4. How would you recover if the interviewer challenged your core requirement or estimate?
  5. Which part of your explanation could be removed without weakening the argument?