Netflix Software Engineer Interview Guide: Problem Solving, Coding, System Design and Culture
Prepare for Netflix software engineer interviews with reported assessment patterns, interview questions, and preparation advice.
On this page
Last reviewed November 15, 2026.
For software engineers at every level.
Built by the Codinzhub team from Netflix interview experiences that candidates shared. Each experience is one person's interview, and Netflix changes its process by team and level, so use this as a map, not a promise.
The 30-second version. A recruiter reaches out or you apply, then a phone screen of one or two rounds (a "problem solving" round and a coding round), then a virtual onsite split into two parts: VO1 is general (coding, system design, behavioral) and VO2 is about the team's own work (domain design, data model, hiring manager). Rounds have different names, such as problem solving, partnership or domain, but underneath they are still coding, system design and behavioral. Netflix weighs culture heavily, so treat the behavioral rounds as seriously as the technical ones.
Round-by-round breakdown
Stage 1: Application, where candidates applied
How it works. In several experiences a recruiter reached out because the candidate had related experience, for example on a data platform. Most teams in these accounts were named: Ads, Member, data platform. One candidate's process took a little over two months from applying to accepting an offer.
Stage 2: Shortlisting, how candidates moved forward
How it works. A recruiter call about your background and the team. Candidates were told which team they were interviewing for, and several said that preparing for that team's domain mattered. One senior candidate asked to interview for Staff and was told an outside L6 maps to L5 at Netflix.
Stage 3: Phone screen
At a glance
- Length: about an hour; one or two rounds, sometimes back to back
- Questions per round: 1 problem with 2 or 3 follow-ups, or 2 to 3 short coding questions
- What was asked: deduplication, sliding windows, shortest paths, a key-value store with expiry, topological sort
- Most common: the key-value store with TTL, and Netflix-flavoured dedup problems
How it works. Usually two parts: a problem solving round, which is about talking through an approach more than writing code, and a coding round. Some candidates called it open book. You may have to write your own test cases. Results came back in one or two days for some candidates and about a week for others.
Questions candidates reported:
| Question (our summary) | Follow-up reported |
|---|---|
| Detect duplicates in a watch history, in three parts | The third part needs bucket sort |
| Deduplicate homepage rows: the first n titles of each row are unique across the page, the rest only within the row | n differs per row; some rows skip the global rule |
| Key-value cache with an expiry time per entry | What to evict when memory is full |
| Fastest broadcast from one city to every other city, given ping times (Dijkstra) | Ping times keep changing |
| Longest substring without repeating characters, told as show names | Optimise with a map of last positions |
| Topological sort | Know the running time; print every valid order |
More questions: Browse Netflix questions.
Stage 4: The virtual onsite
How it works. The onsite is split into VO1 and VO2, sometimes on separate days. VO1 is general: coding, system design and a behavioral round. VO2 depends on the team: a design problem from the team's own work, a data model round, and hiring manager or partner-team conversations. Senior Ads candidates had four or five rounds over two days.
Coding round
At a glance
- Length: about an hour
- Questions per round: 1 problem plus 2 follow-ups
- What was asked: class design with ordering, tree DFS, topological sort with timings
- Most common: the movie billboard or playlist rotation; topological sort with a finish-time twist
Questions candidates reported:
| Question (our summary) | Follow-up reported |
|---|---|
| Rotate movies on a billboard or playlist: each get returns a movie once until all are used; scores can change | Make it thread-safe; scale and store it in production |
| Course prerequisites where each course takes time: when does each one finish? | Handle the edge cases in the timing |
| A tree DFS problem, not from coding-practice | Write and run your own tests |
One candidate said a heap got them stuck and that two sorted maps worked for the rotation problem.
More questions: Browse Netflix questions.
System design round
At a glance
- Length: about an hour
- Questions per round: 1 design, explored point by point
- What was asked: schedulers, subscriptions, ad frequency caps and pacing
- Most common: an ML job scheduler; ad frequency capping for Ads
Questions candidates reported:
| Question (our summary) | What the interviewer focused on |
|---|---|
| Schedule ML jobs (training, batch inference, evaluation) | Requirements per job type, resource allocation, monitoring, crash handling, multi-tenancy |
| A Netflix subscription system | Plans, charging, upgrades and downgrades, retrying failed payments, growth |
| Ad frequency cap | Redis internals, atomic updates, online and offline paths, exactly one ad per cap |
| Ad pacing | Rate limiting, at a high level |
One candidate who handled every question still failed because the design was not finished in time.
More questions: Browse Netflix questions.
Domain and data model rounds (VO2)
At a glance
- Length: about an hour each
- Questions per round: 1 open problem
- What was asked: the team's own systems
- Most common: the Ads data model (advertiser, campaign, ad group, ad)
Questions candidates reported:
- Model the demand side of an ads system: advertiser, campaign, ad group, creative and targeting.
- Manage publisher-specific configuration rules on the supply side.
- A design taken directly from the data platform team's work.
Behavioral, hiring manager and partnership rounds
At a glance
- Length: about an hour each; often two of them
- Questions per round: 3 to 5 questions
- What was asked: collaboration, giving and asking for feedback, conflict, weaknesses, culture
- Most common: giving and receiving feedback
Questions candidates reported: your weaknesses; collaboration and handling conflict; asking for and giving feedback; how your experience matches the job description. A "partnership" round is a manager from a partner team asking behavioral questions.
More questions: Browse Netflix questions.
Stage 5: Decision and offer
How it works. Decisions came about two days after the onsite in one report. Others waited one to two weeks, and one senior candidate heard nothing after a hiring freeze was announced. One candidate who reached the final round was rejected because the last system design ran out of time. Rejections often came with no detailed feedback.
Topics to prioritise
Across these experiences, the most common topics are system design, behavioral and culture, graphs and topological sort, hash maps with sliding windows, and class design with concurrency follow-ups. Netflix problems are often told as Netflix features: rows, billboards, watch history.
The Codinzhub Next-Round Playbook
What Netflix candidates who moved on did. Put together by the Codinzhub team from the Netflix experiences in this guide.
To get past the phone screen
- In the problem solving round, explain your thinking and agree on an approach before coding. Communication was the point.
- Clarify incomplete problems early, for example whether every cache entry has its own expiry.
- Write your own tests, including an error case.
To get past the coding round
- Practise class-design problems with follow-ups on thread safety. Say where the locks go.
- Know the running time of your solution and be ready for variants, such as printing every result.
To get past system design
- Keep the design simple and cover the fundamentals: databases, caches, queues, load balancers.
- Watch the clock. Cover the main points first; the interviewer will pull you back once they have the signal.
- For Ads roles, go deep on Redis, atomic counters and the online and offline paths.
To get past behavioral and hiring manager
- Read the Netflix culture memo several times. Candidates said behavioral rounds do reject people.
- Check that your experience matches most of the job description, and steer your stories toward it.
FAQ
Is the problem solving round different from coding?
Candidates said they felt almost the same. Problem solving puts more weight on talking through the approach.
How long does the process take?
One candidate took a little over two months from applying to accepting an offer.
About this guide
This guide is researched and written by the Codinzhub team. We read every Netflix interview experience we could find, kept only what candidates actually described, and rewrote it in our own words. Questions are our own short summaries; the practice links point to similar problems, not the exact interview questions. Found something out of date? Share your own experience on Codinzhub and we will update the guide.
Keep track of what you’ve learned.