Backend Engineering interview questions and answers

Language-agnostic service engineering: the protocols, boundaries, and failure handling that sit between a coding round and a full system design round.

30 published across 7 topics.

Backend EngineeringA 28-minute read on how the topic fits together, before the questions.Backend engineering fundamentals56 short answers on one page, for revising rather than studying.

masterclass guide

Performance Diagnosis and Optimisation Interviews: A MasterclassHow senior interviewers grade the question every backend loop eventually asks: why is this slow, and how would you find out. Percentiles, queueing, profiling, the slow-query path, tail latency in a fan-out, metastable collapse, autoscaling that browns out, and how to report a finding a business can decide on.104 min

HTTP & REST

4 questions

Methods and status semantics, caching headers, content negotiation, and HTTP/2 and HTTP/3 differences.

mediumConceptDesignScenario

Two clients open the same record, both edit it, and the second save silently overwrites the first. How would you use ETags to turn that lost update into something the client can see and handle?

Return a strong ETag on GET, require If-Match on writes, and reject a stale token with 412 Precondition Failed so a lost update becomes a visible error rather than silent data loss. What makes it correct is that the comparison and the write must be one atomic operation - an UPDATE guarded by the version in its WHERE clause - because a SELECT then UPDATE reintroduces the race.

4 minmid, senior, staff

mediumDesignConcept

Clients are asking for page 4,000 of your /orders collection. How is that endpoint paginated, and what would you change?

Offset pagination gets slower with depth and drifts when rows are inserted mid-traversal, so paginate with a keyset comparison on the sort key plus a unique tiebreaker, hand back an opaque cursor and a Link rel=next header, and price an exact total count separately.

5 minmid, senior, staff

mediumConcept

Which status codes and method semantics do you insist on in a code review, how do the caching headers fit together, and is the API you just described REST?

Get the method's safety, idempotency and cacheability right, use the status code intermediaries and clients act on rather than a generic 200 with an error body, pair Cache-Control freshness with ETag or Last-Modified validators, and be honest that most APIs called REST are resource-shaped HTTP without hypermedia.

6 minentry, mid, senior

hardDesignConcept

Customers need to upload two-gigabyte files to your API over a flaky connection. What does that endpoint look like, and what breaks if you just accept a POST body?

A single large POST fails on the last percent and starts again, and it drags every proxy body limit and idle timeout into your critical path. Either hand out a scoped credential so bytes go straight to storage, or model the upload as a resource with a committed offset the client can resume from.

6 minmid, senior, staff

All 4 in HTTP & REST

Authentication & Authorisation

3 questions

Sessions versus tokens, OAuth 2 and OIDC flows, JWT validation pitfalls, and RBAC versus ABAC.

hardScenarioDesign

A maintenance script updates your permissions table and every user in one tenant can suddenly edit everything. What should have stopped that, and how would you have found out?

Permission data is as privileged as the code that reads it, so a bulk write to it needs the same review, staged rollout and reversibility as a deploy. Derive grants from a source of truth rather than stamping them, assert tenant invariants continuously, and know how long a cached decision keeps the mistake alive.

6 minsenior, staff, lead

hardDesignConcept

Customers want per-project roles rather than one global role. How do you model that, and where in the request does the check happen?

Model permission as a relationship between a subject, a role and a specific resource, not as a role stamped on the user, and push the check down into the data access predicate so a list query and a single-record fetch are authorised by the same rule.

6 minmid, senior, staff, lead

hardConceptDesign

Would you use server-side sessions or JWTs for authenticating your API, and why?

Default to server-side sessions in an opaque HttpOnly cookie, because revocation is a row delete and the browser cannot read the credential. Choose JWTs only when a resource server must validate without calling the issuer, keep them short-lived, and pin the algorithm, issuer, audience and expiry on every verification.

5 minmid, senior, staff

All 3 in Authentication & Authorisation

Microservices

5 questions

Service boundaries, data ownership, saga and outbox patterns, and when a monolith is the right answer.

hardConceptScenarioDesign

A teammate says a saga can just roll everything back if step four fails. What is wrong with that, and what would you tell them to build instead?

A saga has no rollback. Each step already committed in its own database and was visible to everyone, so you can only run new forward transactions that semantically offset it. Compensations are not atomic, some actions cannot be compensated at all, and compensations themselves fail - so a saga needs idempotent handlers and an ordering rule for irreversible steps.

5 minmid, senior, staff

hardDesignConcept

You need to add one field to an event and remove another, and five teams consume it. How does that roll out?

Additive changes with defaults are safe, removals are not, so split every removal into deprecate-then-delete with evidence that nobody reads the field. Register the schema, pick a compatibility mode deliberately, and note that backward compatibility requires consumers to deploy first.

6 minmid, senior, staff, lead

hardDesign

One screen needs data owned by three different services, and it has to render in 200 milliseconds with a filter and a sort applied. How does that request get served?

Fanning out in parallel is fine while every service can answer independently, but a filter or sort spanning owners cannot be composed at all. Either make one service the driver and hydrate its page of identifiers, or build a read model from events and pay for staleness deliberately.

6 minmid, senior, staff, lead

hardScenarioDesign

How do you safely rebuild a corrupted read model in an event-sourced architecture without incurring significant downtime?

Rebuild a corrupted CQRS read model by replaying the event store into a fresh projection, applying upcasters for old event versions, validating parity, and cutting over with a blue-green deployment.

2 minsenior, staff

hardDesignConcept

How do you decide where one service ends and the next begins, and when is a modular monolith the more honest answer?

Draw boundaries on data ownership so exactly one service writes each table, then accept that a workflow crossing them needs a saga of compensating steps and an outbox to make the write and the publish atomic. While the boundaries are still moving, a modular monolith is the honest answer.

7 minmid, senior, staff, lead

All 5 in Microservices

Resilience Patterns

3 questions

Timeouts, retries, idempotency keys, dead-letter queues, and designing for partial failure.

hardScenarioDesign

A dependency that normally answers in 80ms starts taking eight seconds. What in your service reacts, and in what order?

A timeout shorter than the caller's patience is the only defence that works without cooperation. Retries need idempotency, jitter and a budget or they amplify the outage; a breaker must trip on latency as well as errors; and a bulkhead is what stops one slow dependency consuming every thread.

7 minmid, senior, staff

hardScenarioDesign

The trigger for the outage is gone and you have restarted the service, and it collapses again within a minute of taking traffic. Why will it not recover on its own?

The system is being held down by effects it now generates itself: accumulated client retries, an empty cache multiplying database load, and a queue of requests whose callers have already given up. Recovery needs shed load and a controlled ramp, not more capacity.

7 minsenior, staff, lead

hardConceptDesign

A client times out and retries your POST /payments call. How do you make that endpoint idempotent?

Take an idempotency key from the client and store it in a uniquely-constrained row written in the same transaction as the effect, so the dedupe record and the work commit together. Replay the stored response on a repeat, refuse a retry arriving mid-flight, and expire keys past the retry horizon.

5 minmid, senior, staff

All 3 in Resilience Patterns

GraphQL & gRPC

4 questions

Schema design, N+1 resolution, streaming RPC, and choosing a protocol per consumer.

hardScenarioConcept

A gRPC call three services deep hangs, and so does everything waiting on it. What was missing?

A gRPC deadline. gRPC sets none by default, so a client waits as long as the server takes, and every caller up the chain holds a thread or stream with it. Set a deadline at the edge, propagate the remaining budget on every hop, stop work on cancellation, and retry only inside what is left.

5 minmid, senior, staff

hardScenarioDesign

Your public GraphQL endpoint lets one client send a query ten levels deep, and it takes the database down. How do you stop that without breaking legitimate clients?

Estimate GraphQL query cost before execution, multiplying list fields by their enforced page size, and reject anything over budget. Add a depth limit as a backstop, persisted queries for your own apps, execution timeouts, and per-client rate limits counted in cost points, not requests.

5 minmid, senior, staff

hardDesignConcept

The same domain has to be exposed to a mobile app, a partner integration and internal service-to-service traffic. Where does GraphQL fit, where does gRPC, and where does neither?

Choose per consumer: GraphQL where many client shapes make over-fetching the problem, gRPC for internal traffic wanting typed generated stubs and streaming, plain JSON over HTTP for partners who want curl and a CDN. Then batch your resolvers, because the N+1 is structural.

7 minmid, senior, staff

hardScenarioDesign

You move an internal service to gRPC behind the same load balancer your REST services use, and one pod ends up serving most of the traffic. What is happening, and what do you change?

gRPC multiplexes many calls over one long-lived HTTP/2 connection, so a balancer that distributes connections pins each client to one backend for its lifetime. Balance per request with an HTTP/2-aware proxy or client-side subchannels, and bound connection age so scaling out actually moves traffic.

6 minmid, senior, staff

All 4 in GraphQL & gRPC

Distributed Systems Theory

3 questions

Time and ordering, leader election, quorums, idempotence, and classic impossibility results.

hardConcept

One node stops answering. How does the rest of the cluster decide whether it has crashed or is merely slow, and what does that decision cost you?

You cannot distinguish a crashed process from a slow one, so every failure detector is a timeout and every timeout is a guess. Choose which error you can afford: declaring a live node dead risks two writers, waiting risks a stall, and only fencing makes the first survivable.

5 minsenior, staff, lead

hardScenarioDesign

Design a distributed lock manager for a high-frequency trading platform where microsecond latency is critical and stale locks can result in massive financial losses.

Designing an ultra-low latency distributed lock manager, evaluating consensus algorithms, and handling clock drift in high-frequency environments.

3 minsenior, staff

hardConcept

Two servers each write a record stamped with its own clock. Why can you not use those timestamps to decide which write happened first, and what do you use instead?

Wall clocks on separate machines disagree by an unbounded amount and can jump backwards, so their timestamps order events by accident rather than by causality. Logical and vector clocks capture what preceded what, quorums give overlap not order, and a lease is safe only with a fencing token.

7 minsenior, staff, lead

All 3 in Distributed Systems Theory

Backend Scenarios

8 questions

Production situations to diagnose out loud: latency spikes, memory growth, cascading failures, bad deploys.

mediumScenarioCoding

You are dropped into an unfamiliar repository with one failing business rule and an AI assistant that only gives hints. What do you do in the first ten minutes?

Debugging an unfamiliar codebase starts by running the failing test, then reading it as the rule and tracing its call path to the code that computes the wrong value. Let the size of the error narrow the cause, ask the assistant to locate and explain rather than patch, and prove the fix with a test that failed before.

5 minmid, senior

mediumScenarioDesign

Users in three countries book appointments, and one of those countries has just changed when its clocks move. What do you store, and what do you refuse to store?

For a future appointment, store the local wall-clock time plus an IANA time zone ID such as Europe/Lisbon, and treat any UTC instant as a cache you recompute when time zone rules change. Refuse fixed offsets, abbreviations like IST, and server-local times. Past events are stored as UTC instants.

5 minmid, senior

hardScenarioDesign

How do you resolve a complex priority inversion scenario involving chained mutexes and interrupt service routines in a hard real-time system?

Assess the candidate's understanding of advanced priority inversion scenarios in RTOS environments, focusing on chained resources, priority inheritance protocols, and interrupt synchronisation.

3 minsenior, staff

hardScenario

Consumer lag on your orders topic climbs every morning and never clears before the next peak. Work through it.

Lag is the integral of arrival rate minus service rate, so establish which of the two moved and whether the consumer is slow or stalled. Check lag per partition before per group, rule out a rebalance loop, and remember that adding consumers beyond the partition count does nothing.

6 minmid, senior, staff

hardScenarioDesign

How do you optimise UDP packet loss for a high-tickrate competitive shooter?

Analyse approaches to mitigating packet loss in a fast-paced UDP environment, focusing on forward error correction and redundant data.

3 minsenior, staff

hardScenario

Your p99 latency jumped tenfold after a deploy while p50 is unchanged. Diagnose it out loud.

A flat p50 rules out a uniform slowdown, so the tail belongs either to a subset of requests or to a queueing effect. Partition the latency histogram by instance, endpoint and tenant before naming a cause, then confirm with traces, GC logs and pool-wait metrics.

6 minmid, senior, staff

hardScenarioDesign

How do you optimise an FIR filter implementation on a resource-constrained DSP for ultra-low-latency active noise cancellation?

Evaluate the candidate's proficiency in implementing and optimising digital signal processing algorithms on hardware, focusing on FIR/IIR trade-offs, fixed-point arithmetic, and instruction-level parallelism.

2 minsenior, staff

hardScenario

Your database fails over and is accepting connections again within thirty seconds, but your service returns errors for ten minutes. Where is that time going?

The database recovered; your clients did not. Sockets to the old primary hang rather than fail, a cached DNS answer keeps pointing at it, the promoted node refuses writes while it finishes, and then the whole fleet reconnects at once. Attribute the ten minutes to those stages before tuning anything.

6 minmid, senior, staff

All 8 in Backend Scenarios