Glassdoor is the dominant company-reviews and salary-transparency platform — owned by Recruit Holdings (same parent as Indeed). The interview emphasizes user-generated content moderation, salary aggregation, and the unique two-sided dynamics of employees vs employers.
Process
Recruiter screen → 60-minute coding phone (DSA medium) → onsite virtual: 2 coding, 1 system design, 1 craft deep-dive, 1 behavioral. Cycle: 4–6 weeks.
What they actually ask
- Design a moderation system for company reviews (anonymous + authentic) — the tension to resolve is proving a reviewer actually worked somewhere without storing anything that deanonymizes them. Talk through verifying an employment signal (work email domain, a linked profile) at submission time, then stripping identity from the stored record, plus spam scoring and the employer takedown/appeal flow.
- Design salary aggregation with privacy preservation — lead with a minimum sample size: don’t publish a number for a title+location until enough submissions exist, or one person’s pay becomes obvious. Expect follow-ups on outlier trimming, showing a range instead of raw values, and how you stop someone reverse-engineering a coworker’s salary from a bucket of three.
- Design a recommendation system matching users to companies — walk through the features (role, location, pay expectation, past applications), cold-start for brand-new users, and ranking signals like review sentiment and application outcomes. Interviewers probe how you balance relevance against freshness and keep the results from collapsing onto the largest employers.
- Coding: medium DSA, often with search or text-processing framing — expect string parsing, tokenization, prefix/autocomplete matching, and hashing that mirror review search and salary lookup. Drill tries, sliding windows, and frequency maps rather than heavy graph or DP.
- Behavioral: customer focus, ownership, mature product — bring stories about acting on user feedback, owning a messy problem end to end, and making a pragmatic call where reliability outweighed shipping fast. They want signal you can operate calmly on a mature product, not only greenfield work.
Levels and comp (2026)
- SE II: $135K–$170K total
- Senior SE: $200K–$260K
- Staff: $290K–$380K
- Principal: $400K–$520K
Prep priorities
- Be fluent in Java or Scala (the bulk of the codebase) — you’ll write real, idiomatic code in the coding rounds, so know Java collections, streams, and basic concurrency (or the Scala equivalents) cold. Be ready to reason about thread-safety in something like a shared aggregation counter or a cache.
- Understand content moderation and trust/safety patterns — know the building blocks: automated spam/abuse classifiers, human review queues, reputation scoring, and rate limits that blunt review-bombing. A likely prompt: how would you catch a coordinated wave of fake reviews aimed at one employer.
- Brush up on differential privacy / privacy-preserving aggregation — understand why you enforce a minimum group size or add noise before publishing salary stats, and the accuracy-vs-leakage trade-off. Be able to sketch k-anonymity and explain why a raw average over three people leaks individual pay.
Frequently Asked Questions
Is Glassdoor remote-friendly?
Distributed-first. Office in San Francisco. Most engineering roles fully remote within US.
How does Glassdoor compare to Levels.fyi or Blind?
Glassdoor is the broadest. Levels.fyi specializes in tech compensation. Blind is anonymous discussion. Glassdoor pays mid-tier; below tech specialty companies.
What is the engineering culture?
Stable, mature, calm pace. Less expansive than peak years; reliable and humane.
Similar company guides
Prepping for Glassdoor? Put it to work:
