Back to the community

Twitter behavioral interview questions

Researched interview questions, process detail, and difficulty signals for Twitter, compiled by the Primly research team.

6 experiences Difficulty 3.7/5 Technology / Social Media

Data Scientist (Ads Measurement)

virtual · Difficulty 4/5

Candidates report an initial recruiter conversation that confirms domain fit in advertising, experimentation, or causal inference, plus a high level check of stakeholder experience and prior impact. The next step is often a technical screen that blends statistics and product sense, frequently including questions about A/B testing pitfalls, attribution, and how to interpret measurement results for advertisers. Many candidates then complete one or more interviews that resemble a case study where they design an experiment for an ads product change, define success metrics, and describe how they would handle bias from auction dynamics and user heterogeneity. A virtual onsite loop typically includes SQL and data manipulation, modeling or inference depth, and a cross functional interview with Ads Product or Engineering focused on how analyses translate into decisions and roadmap tradeoffs. The final stage often involves a hiring manager debrief or panel discussion that tests senior level ownership, how a candidate scopes ambiguous measurement requests, and how they communicate uncertainty, with end to end timelines commonly around three to five weeks.

  • Design an experiment to evaluate a change to Twitter’s ads ranking that aims to improve advertiser ROI without harming user engagement. What are the primary and guardrail metrics, and how would they be measured?
  • In the presence of ad auctions and budget caps, what are common A/B testing pitfalls on ad platforms, and how would a candidate detect and correct for them?
  • Given a table of impressions, clicks, conversions, spend, and campaign attributes, write SQL to compute incremental lift by segment and identify where results are statistically unstable.
  • Tell about a time a candidate’s analysis conflicted with stakeholder expectations. How was the disagreement handled and what decision was ultimately made?
  • Twitter’s ads business depends on credible measurement to retain advertisers. How would a candidate communicate uncertainty, limitations, and recommended next steps to sales or client teams?

Site Reliability Engineer (SRE)

virtual · Difficulty 3/5

Candidates report starting with a recruiter screen focused on availability, location, work authorization, and a high level review of infrastructure experience, often scheduled within a week of applying or being sourced. The next stage is typically a technical phone or video screen with an engineer that emphasizes Linux fundamentals, networking, scripting, and incident response thinking, usually 45 to 60 minutes. Many candidates then move to a practical assessment, either a live troubleshooting session or an exercise that simulates on call scenarios such as debugging latency, packet loss, or service timeouts in a microservices environment. The virtual onsite loop commonly includes three to five interviews covering systems design at an appropriate scope for entry level, observability and monitoring, and collaboration during outages. A final round or closing conversation may include a hiring manager and sometimes a cross functional partner to assess communication during high pressure incidents and alignment with operational rigor, with overall timelines commonly spanning two to four weeks depending on scheduling.

  • A service behind the Twitter timeline is showing elevated p95 latency and error rates after a deploy. What metrics and logs would be checked first, and how would the rollback versus forward fix decision be made?
  • How would a candidate diagnose intermittent DNS resolution failures affecting only one availability zone, and what data would confirm the root cause?
  • What is the difference between saturation, latency, and errors, and how would those signals be translated into actionable alerts for a high traffic API?
  • Describe an incident where a candidate had to coordinate with developers to restore a degraded system. What communication pattern kept stakeholders aligned?
  • Twitter is sensitive to real time user experience and public visibility during outages. How would a candidate balance speed of mitigation with risk of introducing more instability?

Trust & Safety Engineer

virtual · Difficulty 4/5

Recruiter screen, technical phone, virtual onsite with case study on a content moderation problem, behavioral, and team-fit. T&S team has been heavily restructured post-acquisition.

  • Walk me through building a detection system for abuse.
  • Tell me about handling a high-profile content decision.
  • Describe partnering with policy teams on enforcement.
  • How do you stay credible during platform leadership changes?