How Snap Interviews in 2026: Process, Questions & What They Score
The complete process, reported questions, and what interviewers score.
Snap / Interview snapshot
7
stages from
application to offer
64
reported questions
23
public sources
Independent: SupaCV is not affiliated with or endorsed by Snap. This is one well-documented shape of Snap's interviews, not a script: processes change and vary by role, team, level, and region, so confirm specifics with your recruiter.
On this page
01 / For every Snap candidate
The process, start to finish.
Snap Inc. is the consumer camera and messaging company behind Snapchat, Spotlight, Snap Map and AR Lenses. This playbook covers the track Snap hires for at the highest volume, software engineering, and adds machine learning, data and product management material where public sources documented it; the process shape is shared across those tracks and only the round content changes. Snap publishes an evaluation framework.
Its official careers page "How We Interview" describes competency-based interviewing built on "Craft & Value Competencies", states that "Interviews will include at least two competencies: A Snap value & a craft competency", lists nine Values-Based Competencies grouped under the three company values Kind, Smart and Creative, and tells candidates to answer behavioral questions with S.A.I.L. (Situation, Action, Impact, Learning) rather than STAR.
The nine competency names in the framework list are copied from that page; Kind, Smart and Creative are the group headings, not competencies, so they are not counted as items. Snap does not publish craft competency names, a leveling rubric, a stage list, round counts or durations. A separate official Machine Learning department page lists the four parts of the ML loop (Coding, ML Fundamentals, ML Applied Design, ML System Design) without durations.
Every number below comes from coaching guides plus a few first-hand Blind threads and is given as a range. Overall shape: recruiter screen, one technical or functional screen, then a loop of roughly four to six rounds of 45 to 60 minutes, run in one day or spread out (Snap's page confirms only that both happen). Coding rounds dominate for engineers, with system design added from mid-level upward.
Sources split on behavioral assessment: Exponent's ML guide and interviewing.io describe it as woven into every round for 10 to 15 minutes with no standalone interview, while Prepfully, TechPrep, AlgoCademy and Interview Kickstart list a dedicated behavioral or culture round. The loop is centralized: a debrief decides hire or no-hire independent of team, team matching calls follow, and the offer is tied to a team. Sources disagree on whether team matching completes before or after the offer.
Stage 1: Application or recruiter outreach
varies
Applications go through the Snap careers portal, or a recruiter reaches out. No source states a response time. Snap also runs a separate Emerging Talent pipeline for interns and new grads plus Snap Academies for community college students, but the official page publishes no application windows or interview steps for it.
Stage 2: Recruiter screen
20 to 60 minutes
Background, resume walk-through, motivation for Snap and compensation expectations. Reported lengths span 20 to 30 minutes (CodingInterview.com), 30 minutes (Exponent PM guide), 30 to 45 minutes (TechPrep), 30 to 60 minutes (Exponent blog and ML guide) and a full hour (interviewing.io). TechPrep says recruiters often ask about experience with stacks like Java, Go or C++, while Exponent's blog says most coding roles do not require a specific language. CodingInterview.com says recruiters assess your knowledge of Snap's ecosystem (Snapchat, Bitmoji, Spectacles, Lens Studio). No source says recruiters ask how you personally use the app.
Stage 3: Technical or functional screen
45 to 60 minutes for engineers; 30 to 60 minutes for the PM hiring manager call
For engineers, a remote coding session in a shared editor, commonly HackerRank over a video call, with one or two problems plus a short project deep dive; Exponent's blog reports roughly 10 to 20 minutes of behavioral or background questions before the coding. Exponent's SWE guide says code must run, not pseudocode. For PM candidates this stage is a 30 or 60 minute conversation with the hiring manager covering communication and past work. Both CodingInterview.com guides describe a role-dependent online assessment before or instead of this; Snap's own page does not confirm one, so treat it as role dependent.
Stage 4: Onsite or virtual loop
4 to 6 rounds of 45 to 60 minutes, about 4 to 6 hours in total
Run in one day or spread out; Snap's own page states only that both happen and Prepfully and TechPrep say candidates can split it over two days. Engineering loops are reported as two to four coding rounds, one system design round for mid-level and above, either a dedicated behavioral or culture round (Prepfully, TechPrep, AlgoCademy, Interview Kickstart) or behavioral questions spread across every round with no standalone interview (Exponent ML guide, interviewing.io), and an informal or lunch chat with another employee (Exponent's SWE guide says a non-technical employee, Prepfully says another engineer). Prepfully adds that the design round may include what you like and dislike about the product. interviewing.io reports a heavier six-hour loop of four coding hours, one system design hour and a 30 to 60 minute Q&A. For ML roles Snap's official department page lists four parts, Coding, ML Fundamentals, ML Applied Design and ML System Design, with no durations; Exponent reports two 60-minute coding rounds, one to two ML fundamentals rounds and applied ML or ML system design rounds. PM loops are five 45-minute interviews with three PMs, one engineer and one non-product person; Exponent says PMs are assessed on communication, product intuition, analytics and relevant experience. One Blind commenter describes the technical questions as harder and less typical than other large-company LeetCode style questions and the behavioral evaluation as stricter than usual.
Stage 5: Debrief and hiring decision
varies; one candidate reports about 2 business days
interviewing.io and a Snap employee on Blind describe a centralized process: interviewers debrief and decide hire or no-hire irrespective of team. The same Blind employee says one weak interview does not automatically disqualify a candidate; it depends on how the other rounds go. One Blind poster who passed an L4 backend loop reports finishing the onsite on a Thursday and getting the committee decision in about two business days. Exponent's PM guide says candidates hear back within a few days after each round. No official statement on committee or debrief mechanics exists.
Stage 6: Hiring manager conversations and team matching
varies
Candidates generally do not interview for one team; after the loop they are allocated team matching calls that fit their interests, and the Blind employee adds that the manager has to want you. TechPrep describes a brief final hiring manager conversation sometimes folded into the onsite; AlgoCademy and two Blind threads place team matching calls after a passed onsite; interviewing.io places team matching after accepting an offer and says it can take several weeks or longer across multiple hiring manager calls. A Blind poster interviewing for two named teams, each hiring manager having said which team they hire for, shows the sequence can vary. Length depends on which teams have headcount.
Sources disagree on whether team matching happens before the offer (Blind, AlgoCademy) or after offer acceptance (interviewing.io). No official statement.
Stage 7: Offer
varies
A Blind thread's original poster states the offer is tied to a team. End to end, guides report 3 to 4 weeks (Exponent blog), 3 to 5 weeks (CodingInterview.com), 3 to 6 weeks (TechPrep), as little as 3 weeks (Exponent SWE guide) and around one month (Interview Kickstart). Expect the range to widen with seniority and with team matching. A commenter in the same Blind thread says a recruiter told them feedback stays valid for a year, but the thread does not make clear whether that describes Snap, so it is not relied on.
Reported total process length ranges from 3 to 6 weeks across sources; treat any single figure as approximate.
Reported interview time, one scale in minutes
Where the minutes go
Onsite or virtual loop is drawn as its total: 6 rounds of ~60 min each.
- Recruiter screen60 min
- Technical or functional screen60 min
- Onsite or virtual loop~360 min in total
Not drawn: Application or recruiter outreach, Debrief and hiring decision, Hiring manager conversations and team matching and Offer, whose durations are not given in minutes.
02 / What Snap scores
Craft & Value Competencies
01
Courage
02
Empathy
03
Instills Trust
04
Action Oriented
05
Decision Quality
06
Strategic Mind
07
Manages Ambiguity
08
Cultivates Innovation
09
Insatiable Learning
From Snap's official "How We Interview" page (https://careers.snap.com/how-we-interview). Snap describes competency-based interviewing using "Craft & Value Competencies" and states: "Interviews will include at least two competencies: A Snap value & a craft competency." Values-Based Competencies are always assessed with behavioral questions; Craft Competencies "are assessed using either behavioral questions or demonstrations of skill." The nine items above are the Values-Based Competencies exactly as listed.
The page groups them under three bold headings that are Snap's company values, introduced by "Our values in action guide how we act, work, and show up": Kind (Courage, Empathy, Instills Trust), Smart (Action Oriented, Decision Quality, Strategic Mind), Creative (Manages Ambiguity, Cultivates Innovation, Insatiable Learning). Kind, Smart and Creative are headings rather than competencies, so they are not counted as framework items; Snap's Culture page independently confirms them as the company values.
Read the rest of the framework notesShow less
Snap does not publish the Craft Competency names, which are role and discipline specific, so none are listed. Snap also publishes an answer structure for behavioral questions, S.A.I.L.: Situation ("What's the background? What problem needed to be solved?"), Action ("What was your role? What actions did you take?"), Impact ("What were the results? What impact did it have?"), Learning ("What did you take away from this experience?").
S.A.I.L. is an answer format, not a competency, and is not in the framework list. Several coaching guides (Exponent's SWE guide, TechPrep, AlgoCademy) still recommend STAR and Exponent's PM guide recommends a three-point "Triangle" structure; the official page replaces Result with Impact and adds an explicit Learning step. Snap Engineering separately publishes eight engineering values at https://eng.snap.com/values (Customer Focus, Execution, Craftsmanship, Creativity, Accountability, Empathy, Integrity, Audacity).
That page says nothing about interviews and the How We Interview page does not reference it, so those values are not part of the interview framework list; Exponent's ML guide repeats them as culture context, and Prepfully reports a behavioral prompt about demonstrating audacity.
Interview mechanics stated on the official page: a "camera on" policy for virtual interviews; "unless specifically told you are allowed, external resources, including AI, aren't allowed during interviews"; time is always left for candidate questions; and "Sometimes interviews are grouped into one day; other times, they're spread out." On dress the page says the team tends to dress casually and to "wear whatever makes you feel comfortable and confident"; that is guidance, not a rule.
03 / Coach's notes
Coach's tips.
01
Answer behavioral questions with S.A.I.L., not STAR. Snap's own careers page prescribes Situation, Action, Impact, Learning. The two steps candidates skip are Impact, which Snap wants quantified rather than described, and Learning, which is a separate beat about what you took away. Rehearsing STAR stories and stopping at the Result leaves a quarter of the rubric unanswered.
02
Bring stories that each carry two competencies at once. Snap states that every interview covers at least one Snap value plus one craft competency, so build a small set of stories and label each with the competency it demonstrates (for example Manages Ambiguity, Instills Trust, Decision Quality) and the technical craft it shows. A story that only shows technical skill leaves half the interview unscored.
03
Know the product before you interview. No source says recruiters quiz you on your own usage, but CodingInterview.com says recruiters assess your knowledge of Snap's ecosystem (Snapchat, Bitmoji, Spectacles, Lens Studio), Prepfully reports what you like and dislike about the product being asked in the design round, Interview Kickstart reports what you would change about Snapchat, Exponent tells candidates to download and use Snapchat, and system design prompts are drawn from Snap's own surfaces (Stories, Spotlight, Snap Map). A Blind commenter reports a recruiter stressing passion for Snap's business and values and behavioral answers being judged more strictly than usual.
04
For system design, prepare the ephemeral and real-time media cases specifically. Snap's distinguishing constraints are disappearing content, low-latency camera and AR paths, and mobile-first delivery. Get comfortable talking about TTL and deletion semantics, media upload pipelines and CDN distribution, and fan-out for Stories, rather than defaulting to a generic CRUD-plus-cache diagram. TechPrep and CodingInterview.com report system design as applying from mid-level or senior upward.
05
Write code that compiles and runs. Exponent's SWE guide reports that Snap interviewers are sticklers for code that runs and that pseudocode is not accepted; the screen is a shared editor, commonly HackerRank, and the expectation is working code with the edge cases handled.
06
Follow the stated interview mechanics. Snap's page requires cameras on for virtual interviews and prohibits external resources and AI tools unless the interviewer explicitly permits them. It says the team tends to dress casually and to wear whatever makes you comfortable. It also allocates time for your questions in every round, so arrive with specific ones.
07
Do not assume you are interviewing for one team. interviewing.io and a Snap employee on Blind describe a centralized loop where a debrief decides hire or no-hire irrespective of team and team matching calls follow, so your loop interviewers may not be your future teammates and your stories should generalize beyond one domain. The same employee says one weak round is not automatically disqualifying if the others go well. Budget time for matching: interviewing.io says it can take several weeks.
04 / The research
Where this page comes from.
Compiled from public sources across 13 sites. Snap's own page is among them: careers.snap.com/how-we-interview.
- sites
- 13
- coaching and prep guides
- 14
- Snap's own pages
- 5
- candidate reports and forums
- 4
Source composition
Where the research comes from
Grouped by the kind of site that published each source: Snap's own domain, candidate forums, and everything else.
- Coaching and prep guides14
- Snap's own pages5
- Candidate reports and forums4
05 / The questions
Reported questions, by category.
Showing questions for
Roles are grouped by the category each question fits, not by the source.
Coding
18 questions
Reverse a linked list in place.
Iterative three-pointer reversal of a singly linked list in O(n) time and O(1) space.
Copy a linked list with random pointers.
Deep copy with a hash map from original to clone nodes, or the interleaved-node O(1) extra space trick.
Design and implement an LRU cache.
Hash map plus doubly linked list for O(1) get and put with capacity eviction.
Find the minimum window substring of s containing all characters of t.
Sliding window over s with character counts; tests expand and shrink two-pointer logic.
Word ladder: find the shortest transformation sequence between two words.
BFS over words that differ by one letter; shortest path in an unweighted graph.
Find the longest palindromic substring in a string.
Expand around center in O(n^2) or Manacher's algorithm; tests string DP and odd and even centers.
Serialize and deserialize a binary tree.
Preorder traversal with null markers; tests recursion and round-trip encoding design.
Build a binary tree from an array.
Level-order array to tree construction using index arithmetic or a queue.
Given a matrix of integers, find the maximum sum of a k by k submatrix.
2D prefix sums give O(1) window sums after O(mn) precompute; Exponent lists 2D arrays as a Snap topic.
Trapping rain water: compute how much water collects between buildings of varying height.
Two-pointer or prefix max arrays; water at i is min(left max, right max) minus height.
Find the shortest path through an M by N matrix of zeros and ones.
BFS on a grid from source to destination with visited tracking; tests neighbor iteration.
Determine whether a given integer is a power of two.
Bit trick: n and (n-1) equals 0 for powers of two; AlgoCademy separately reports counting set bits.
Implement a basic calculator that parses and evaluates an expression string.
Stack-based expression parsing with operator precedence; Exponent lists it as a topic, not a word-for-word prompt.
Given a string of digits and a target value, insert +, - and * operators so the expression evaluates to the target.
Backtracking over operator placement, tracking running value and last operand for multiplication.
Given exchange ratios between symbols, compute the conversion ratio between any two.
Weighted graph traversal multiplying edge ratios along the path; weighted union-find also works.
Given bus routes and stops, find the minimum number of transfers between two stops, or -1 if impossible.
BFS over routes rather than stops, so each hop between routes counts one transfer.
Remove duplicates that appear within k positions of each other in an array.
Sliding window with a hash set of the last k values; tests window maintenance.
Parse an XML document into a tree structure.
Stack-based tag parsing to build a tree of nodes; tests parsing and recursion.
Reported problems, listed so you know what to expect. Practice them in your own editor or on the platform you prefer; SupaCV's practice mode coaches how you talk through them.
System Design
15 questions
Design the backend for Snapchat Stories.
Fan-out for followers, media storage and 24 hour expiry; Interview Kickstart reports the Instagram Stories variant.
Design Snapchat's ephemeral messaging system, including how messages expire.
TTL and deletion semantics, delivery state and storage lifecycle; TechPrep lists it under low-level design.
Design Snap Map, showing friends' locations in near real time.
Geospatial indexing, location update fan-out, privacy controls and update frequency trade-offs.
Design the Spotlight recommendation feed.
Candidate retrieval, ranking and prefetch for a scrollable video feed; Exponent and Prepfully report generic variants.
Design a scalable push notification service for Snaps, messages and app events.
Fan-out to devices, token registry, retries and rate limits across APNs and FCM.
Design a chat or messaging application at Snapchat scale.
Delivery, ordering, presence and persistence for millions of DAU; TechPrep frames it as WhatsApp or Messenger.
Design the media upload pipeline and global CDN distribution for Snaps and video.
Chunked upload, transcoding and CDN edge caching; Prepfully and Exponent report generic video variants.
Design an ad server.
Ad selection, targeting, auction latency and impression tracking; interviewing.io lists it as a typical topic.
Design an API rate limiter.
Token bucket versus sliding window counters, then distributed enforcement across nodes.
Design a URL shortener.
Key generation, collision handling and read-heavy redirect caching.
Design a real-time content moderation system for user-generated content.
Streaming classification, human review queues and latency versus accuracy trade-offs.
Design a 'people you may know' system.
Friend-of-friend graph traversal, candidate scoring and offline precomputation.
Design a system to detect and prevent fraudulent activity.
Real-time signals, rule plus model scoring and feedback loops from confirmed fraud.
Design a photo sharing service.
Upload, storage, feed and CDN; both sources present it as an example topic rather than a reported prompt.
Design Tic-Tac-Toe.
Object-oriented design of board, players and win detection; listed under low-level design.
Reported problems, listed so you know what to expect. Practice them in your own editor or on the platform you prefer; SupaCV's practice mode coaches how you talk through them.
Product Sense
8 questions
How would you improve the Snapchat Discover page?
How would you launch Snapchat in a new country, for example Indonesia?
How would you measure the success of Snap Map?
How would you improve the Spotlight feature at Snapchat?
Why do you think Instagram copied Snapchat Stories?
How is Snap different from other companies in the space?
Design a social media application for introverts.
How many windows are there in Tehran?
Machine Learning and Data
8 questions
How do you choose evaluation metrics for a model, and how do you handle unbalanced data?
Precision, recall, AUC selection plus resampling or class weights; Exponent lists these as topics, not as word-for-word questions.
Explain overfitting, how you detect and prevent it, and how you pick an optimizer.
Train versus validation gap, regularization, early stopping and optimizer choice; listed as topics by Exponent.
Design a ranking or recommendation model for a Snap surface, for example ad ranking or feed ranking.
Feature design, objective choice, offline and online evaluation; Exponent's blog reports an Instagram ranking model.
Implement k-nearest neighbors and k-means clustering.
Distance computation and iterative centroid updates; tests writing ML algorithms from scratch.
Explain how LLMs might be susceptible to adversarial attacks.
Prompt injection, jailbreaks and adversarial inputs; tests ML security awareness.
Given conversion rates for groups A and B, determine whether the difference is statistically significant.
Two-proportion z-test or chi-square test, p-values and sample size.
How would you use data to help Snap engineering improve phone camera speed?
Metric definition, instrumentation and segmentation by device; open-ended analytics prompt.
Explain the difference between a data warehouse and an operational database.
OLAP versus OLTP workloads, schema design and query patterns; reported for data engineering roles.
Reported problems, listed so you know what to expect. Practice them in your own editor or on the platform you prefer; SupaCV's practice mode coaches how you talk through them.
Machine Learning and Data
1 question
Write SQL to compare snaps sent against snaps opened, or to find reciprocal friendships.
Joins, self-joins and aggregation; Exponent also reports DAU queries and an ad revenue calculation.
Reported problems, listed so you know what to expect. Practice them in your own editor or on the platform you prefer; SupaCV's practice mode coaches how you talk through them.
Every role
Behavioral and Values
14 questions
Why do you want to work at Snap?
Walk me through your resume.
Tell me about a time when you had to solve a difficult technical problem.
Tell me about a time you turned a creative idea into a practical solution.
How do you handle feedback and rapid iteration?
Describe a project where you challenged assumptions to improve the user experience.
Tell me about a project you are proud of and the impact it had.
What would you change about Snapchat?
Describe a situation where you had to work with a difficult team member.
Describe a situation where you gave difficult feedback to a teammate.
How do you stay current with new technologies, and how do you approach learning one you have never used?
Talk about a time you demonstrated audacity.
Tell me about a time you failed. What did you learn?
How do you plan your approach when working with an engineering team?
Your next step
Interviewing at Snap?
Make sure your resume gets you to the first round.
Practicing for Snap against your resume is in Interview Prep on Career.
Common questions
Quick answers.
How many stages are in Snap's interview process?
Snap has 7 stages, set out in The process.
What framework does Snap use to evaluate candidates?
Snap evaluates candidates against Craft & Value Competencies, set out in What Snap scores.
What kinds of questions does Snap ask?
Snap's questions fall into 5 categories, grouped in The questions.
How reliable is this Snap interview playbook?
It is compiled from public sources, and the evidence is in The research.
Sources (23)
Every public source this page was compiled from. Links open in a new tab.
13 sites, by sources
- careers.snap.com4
- tryexponent.com4
- teamblind.com4
- codinginterview.com2
- eng.snap.com1
- interviewing.io1
- prepfully.com1
- interviewkickstart.com1
- techprep.app1
- systemdesignhandbook.com1
- algocademy.com1
- joinleland.com1
- en.wikipedia.org1
- 01careers.snap.com/how-we-interview
- 02careers.snap.com/culture
- 03careers.snap.com/emerging-talent
- 04careers.snap.com/departments/machine-learning
- 05eng.snap.com/values
- 06tryexponent.com/blog/snap-interview-process
- 07tryexponent.com/guides/snap-software-engineer-interview
- 08tryexponent.com/guides/snap-machine-learning-engineer-interview
- 09tryexponent.com/guides/snap-pm-interview
- 10interviewing.io/snap-interview-questions
- 11prepfully.com/interview-guides/snap-software-engineer
- 12interviewkickstart.com/blogs/interview-questions/snap-interview-questions
- 13techprep.app/blog/snap-inc-interview-process
- 14codinginterview.com/guide/snap-interview-process
- 15codinginterview.com/guide/snap-interview
- 16systemdesignhandbook.com/guides/snap-system-design-interview
- 17algocademy.com/blog/snap-inc-technical-interview-prep-a-comprehensive…
- 18joinleland.com/library/a/how-to-prepare-for-snap-product-management-i…
- 19teamblind.com/post/snap-interview-difficulty-tgrlfxs4
- 20teamblind.com/post/Interviewing-with-multiple-Snap-teams-34tsUSBk
- 21teamblind.com/post/Snap-team-matching-and-possible-TC-range-w5HB54dT
- 22teamblind.com/post/Snap-Team-Match-hqhBsi2t
- 23en.wikipedia.org/wiki/Snap_Inc.
Related reading
- STAR, XYZ, or CAR: Which Resume Bullet Formula Actually Works?STAR, XYZ and CAR all collapse to the same bullet in practice, why "quantify everything" is more contested than it sounds, and the one modern use of STAR that actually earns its keep.
- Jake's Resume: Why One LaTeX Template Took Over Tech HiringThe history behind the internet's most-copied resume template, why its formatting choices actually work, and how to get the same layout without touching LaTeX.
How similar companies interview
Compare Snap's process with other Consumer Tech & Media companies in this playbook: