Zillow is the dominant US real estate platform — 200M+ monthly visitors. The interview is data-heavy with focus on Zestimate (home valuation ML), search at scale, and the unique challenges of real estate data (slow-changing, geographically distributed).
Process
Recruiter screen → 60-minute coding phone (DSA medium) → onsite virtual: 2 coding, 1 system design, 1 craft deep-dive, 1 behavioral. Cycle: 3–4 weeks.
What they actually ask
- Design Zestimate — automated home valuation at national scale. Focus on the batch-versus-online split: features like square footage, tax history, and comparable sales are precomputed on a schedule, while the serving layer returns a cached estimate in milliseconds. Interviewers probe how you handle stale data, expose a confidence interval, and re-score 100M+ homes when a model version ships or a neighborhood’s recent sales move.
- Design real estate search with geographic filters. Expect to reason about bounding-box and radius queries, map-tile aggregation for the “search as you pan the map” experience, and spatial indexing (geohash or R-tree) so a filtered map view stays fast. A common follow-up combines a text filter (3-bed, under $500K) with a moving map viewport and asks how you keep both cheap.
- Design real-time MLS integration. MLS feeds arrive from hundreds of regional systems in different formats and cadences, so the design centers on ingestion, normalization to a common schema, and deduping the same listing that shows up in multiple feeds. Be ready to talk about keeping listings fresh (change detection, TTLs), reconciling conflicting updates, and whether a relational or NoSQL store fits the write and query pattern.
- Coding: medium DSA, often with geo or ML framing — for example, cluster nearby listings, merge overlapping price ranges, or find the k nearest homes to a point. The solutions are standard (heaps, hash maps, interval merging), but you’ll be expected to state the time and space complexity and justify the data structure for the geo twist.
- Behavioral: customer focus, ownership, working with industry data. Come with stories about shipping a feature that moved a real user metric and about owning an ambiguous, messy data problem end to end. Zillow leans on customer-obsession language, so tie your decisions back to the buyer, seller, or agent you were actually serving.
Levels and comp (2026)
- SE II: $145K–$180K total
- Senior SE: $215K–$285K
- Staff: $310K–$410K
- Principal: $430K–$570K
Prep priorities
- Be fluent in Java (legacy core) and Python (ML services). Core platform services and much of the older search stack are Java, while valuation and recommendation code is Python — find out which one your target team uses and be ready to write idiomatic code in it, not pseudocode.
- Understand geographic data and MLS standards. Know the basics of lat/long, bounding boxes, and geohashing, plus what an MLS is and why listing data is fragmented across regional feeds — this context shows up in both the search and integration design rounds.
- Brush up on ML for valuation and recommendations. You don’t need to derive gradient descent, but be able to talk through the features you’d use for a home-price model, how you’d measure its error (for example, median absolute percentage error), and how ranking works for recommended listings.
Frequently Asked Questions
Is Zillow remote-friendly?
“Cloud HQ” — distributed-first. Some hubs in Seattle, Bay Area, NYC. Many engineering roles fully remote within US.
How does Zillow compare to Redfin or Realtor.com?
Zillow has the largest brand and visitor base. Redfin is more vertical-integrated. Realtor.com (Move/News Corp) competes on agent listings. Zillow comp is mid-tier; below FAANG.
What is the engineering culture?
Mature, distributed-first, calm pace. Strong work-life balance. Less hype than peak years; reliable.
Similar company guides
Prepping for Zillow? Put it to work:
