Engineering IC interview prep.

A senior SWE IC at a consumer-internet firm operates at planet-scale (hundreds of millions to billions of users), where bps of latency move DAU, bps of conversion move revenue, and the A/B platform is always running.

What interviewers look for

  • Can the candidate run a senior consumer-scale design loop end-to-end: clarify, estimate at hundreds-of-millions, define API + data + architecture, deep-dive hot keys + fan-out + caching - not start drawing boxes?
  • Do they pick consistency + scale primitives to consumer use cases (eventual consistency for feeds, strong for money / auth, fan-out on write vs read, multi-region edge) - and explain the tradeoff?
  • Can they reason about the A/B + growth stack: hypothesis → exposure → metrics + guardrails → ramp; understand consumer experimentation is always-on?
  • Do they show senior behavioral signals: technical influence, disagreement + commit, ownership of failure, communication across teams - not 'I coded the feature'?
  • Can they debug production at consumer scale: read the four golden signals + the business funnel, form hypotheses, isolate the change, validate the fix, prevent recurrence?
  • Are they ecosystem-aware: recommendation / ranking, ad serving, feature flagging, growth funnels, mobile + web differences - the consumer-internet stack characteristics?

Behavioural questions to expect

  1. Walk me through your CV.

    What it tests: Story coherence + genuine fit for consumer-internet IC work. Teams want evidence of progressive technical scope (feature → system → cross-team architecture), consumer-scale exposure, and technical influence - not just years.

  2. Tell me about your most impactful technical project.

    What it tests: Depth + ownership + willingness to defend a technical choice. Tests whether the candidate frames problem → approach (with tradeoffs) → result (quantified) → lessons, not just describes a feature.

  3. Tell me about a weakness, a failure, or feedback you've received and worked on.

    What it tests: Self-awareness + production discipline. Cross-role canonical. Fake weaknesses downgrade immediately. Senior consumer-SWE mistakes (the shipped bug that broke for a hot tenant / segment, the A/B that hid a guardrail regression, the over-engineered design) shape teams.

  4. Why consumer internet engineering - and why this firm specifically vs enterprise SaaS or infra?

    What it tests: Authentic fit for the consumer-scale + A/B-driven + product-led seat: planet-scale, fast feedback loops, direct user impact, recommendation / growth / monetization integration. Tests whether the candidate is drawn to consumer-internet problems specifically.

  5. Which team or technical area would you want to work on, and why?

    What it tests: Genuine fit + grasp of how consumer-internet SWE areas differ. Tests whether the candidate has a reasoned preference (core product backend / mobile / growth / recommendation + ranking / ads / search / payments / infra) rather than 'wherever'.

  6. Why this firm?

    What it tests: Whether the candidate has done the homework. Bar: firm-specific evidence from the product, eng culture, stack, scale, recent launches - not generic 'great tech'.

  7. How would you describe this firm's engineering organisation + architecture in your own words?

    What it tests: Whether the candidate has internalized HOW the firm builds at consumer scale - org shape, stack, architecture, A/B + growth practice - not just that it 'has engineers'. Tests whether they've read the eng blog.

  8. How does engineering actually drive value at a consumer-internet firm?

    What it tests: Whether the candidate understands consumer-internet engineering economics: shipping features that move DAU + retention + monetization, A/B-validated; reliability at planet scale; performance / latency moves engagement; platform + infra scale the feature teams.

Technical concepts to master

CAP + consistency at consumer scale

CAP at consumer scale
Under Partition: CP (consistency) for money / auth / single-source-of-truth; AP (availability) for feeds / counters / search / recommendations. Latency is the often-missed third axis (PACELC).
Eventual consistency + read-your-writes
Eventual: reads may be stale; converges if writes stop. Read-your-writes: user always sees their own writes back (via sticky routing or local cache).
Fan-out on write vs read
Fan-out on write: when user posts, write to all followers' feed lists (fast read, heavy write, bad for celebrities). Fan-out on read: query followers' posts at read time (heavy read, simple write, scales to celebrity).
Idempotency at scale
Retries + at-least-once delivery + dedup; the only way to be safe in a distributed consumer system.

Sharding + replication + hot keys (consumer)

Sharding by user / content
Most consumer data shards by user_id (for user-owned data) or content_id (for shared content); pick the key to avoid hot shards + cross-shard joins.
The celebrity / hot-key problem
When one user (a celebrity, a viral post) generates disproportionate traffic to one shard, overwhelming it.
Multi-region + edge serving
Geo-distributed users need data near them; reads served from edge / regional replicas; writes either anchored to home region or replicated multi-master.
Cache + invalidation at consumer scale
Multiple cache layers (CDN, edge, local app cache); TTL for predictable staleness, event-driven (CDC / pub-sub) for fast invalidation when freshness matters.

A/B testing + growth engineering

A/B platform architecture
Hypothesis → exposure assignment (deterministic by user_id + experiment_id) → metric collection → statistical analysis → ramp / kill decision.
Sample size + statistical power
To detect a 1% relative lift on a 5% conversion baseline with 80% power + 95% confidence: ~750K users per arm. Larger lifts need fewer; smaller need more.
Primary + guardrail metrics
Primary: the metric the experiment is trying to move (engagement, conversion). Guardrails: metrics that must NOT regress (revenue, retention, latency, error rate); a primary lift that breaks a guardrail is a net negative.
Growth funnel + experimentation
End-to-end user journey (acquisition → activation → engagement → retention → monetization → referral); each stage instrumented + A/B-tested for compounding lift.

Observability + golden signals (consumer)

The four golden signals
Latency, Traffic, Errors, Saturation - the four metrics every service emits + dashboards from day 1. Reused from the canonical SRE reference.
Business funnel + metric instrumentation
Beyond golden signals: instrument the business funnel (acquisition → activation → engagement → conversion → retention) so regressions are catchable + attributable.
SLI / SLO / error budget
SLI: measured metric (P99 latency, success rate). SLO: target (e.g. <300ms P99, 99.9% success). Error budget: 1 - SLO; how much unreliability you can spend before slowing feature work.
Blast radius + canary / staged rollout
How much of the user base is exposed if a change is bad; minimized by canary (small %) + staged rollout (region by region) + feature flags + kill switches.

Practical drills

  • A consumer feed service has 500M DAU. Each user opens the app 5x / day on average; each feed request returns 20 items + rendering takes 1MB of media. P99 latency target 300ms. (a) Average + peak QPS. (b) Daily bandwidth. (c) Storage (assume each item is 500B + media is CDN-cached). (d) Rough estimate of feed-serving infrastructure (servers, cache, CDN).
  • Design a personalized recommendation feed for this firm's consumer app at 200M DAU. Walk me through it.
  • this firm's DAU has dropped 3% over the last 48 hours, primarily in mobile. Walk me through how you'd diagnose + fix.

Smart-question anchors

  • Team + scope - the team's surface area, what the role would specifically own in 6-12 months
  • Stack + architecture - the current stack, recent architecture changes, where the eng team thinks it's heading
  • A/B + experimentation culture - cadence, platform, OKR integration, how decisions are made on experiments
  • On-call + reliability - on-call rotation, SLO + error-budget approach, recent incident learnings
  • Career ladder + growth - what differentiates Mid / Senior / Staff at this firm, mentoring, IC track

Related roles

Sourced from

Ready to Generate Your Own Prep?

Drop your CV and a job description on the home page. A couple of minutes later you get a report with everything you need to land the job.

Generate my first prep