Technical interview guide

NoSQL, Graph & Key-Value Data Stores

When a relational database isn't the right fit — document, key-value, graph, and vector stores, and how to choose between them.

Read: 23 min
Practice MCQs: 25
Interview QA: 25
Edition: v3
Editorial status: Reviewed
Relevant for: AI EngineerBackend EngineerBusiness Intelligence DeveloperCloud ArchitectComputational BiologistCybersecurity EngineerData AnalystData ArchitectData EngineerData ScientistForward Deployed Engineer (FDE)Full-Stack DeveloperSoftware EngineerSolutions ArchitectTechnical Product Manager

Scope: SQL principles with PostgreSQL 18 examples; vendor-specific behavior must be verified.

Interview QA

Treat each question like a live interview question: answer out loud first (structure, assumptions, tradeoffs), then open the model answer to spot gaps and rehearse a tighter follow-up.

Curated: 2026-10-05 · Written: 2026-10-05 · Reviewed: 2026-10-03

QA-1

Explain what NoSQL means and why the term is not very useful.

QA-2

How do you decide whether to embed or reference in a document model?

QA-3

Explain how wide-column data modelling works and why it inverts relational practice.

QA-4

When is a graph database the right choice, and when is it not?

QA-5

What would you use Redis for, and what would you not use it for?

QA-6

A team wants to move from PostgreSQL to MongoDB for schema flexibility. How do you respond?

QA-7

How would you design a DynamoDB table for a known set of access patterns?

QA-8

What does 'schemaless' actually mean, and what does it cost?

QA-9

How would you decide between polyglot persistence and one database?

QA-10

How do tombstones affect read performance and compaction in distributed wide-column stores like Cassandra, and how do you design against tombstone-related latency spikes?

QA-11

What is a hot partition and how would you design around one?

QA-12

How would you model a social feed, and which store would you use?

QA-13

How do you handle the loss of foreign keys in a store that has none?

QA-14

Explain when denormalisation is the design rather than a compromise.

QA-15

When modeling variable-depth relationships like fraud rings or recommendation paths, what mechanical advantage does index-free adjacency offer over recursive SQL joins?

QA-16

How would you migrate data from a relational database to a document store safely?

QA-17

Explain the trade-offs of storing JSON in a relational database versus a document store.

QA-18

How would you implement a cache correctly in front of a database?

QA-19

What would make you recommend against a NoSQL store for a new project?

QA-20

Explain eventual consistency and how it affects application design.

QA-21

How would you evaluate whether an existing NoSQL choice was a mistake?

QA-22

Explain how you would model a chat application's message storage.

QA-23

What operational differences should a team expect when adopting a distributed datastore?

QA-24

How would you explain the SQL versus NoSQL decision to a non-technical stakeholder?

QA-25