Snowflake Interview Questions — Page 2

Snowflake·Software Engineer·Technical Phone Screen

Apr 2026

Snowflake SWE interview that went through a string matching problem with two progressively nastier follow-ups. The core question was approachable but the stream-processing variant at the end is where things got real.

Given an ordered list of ingredients and a list of recipes (each recipe is an ordered sequence of ingredients), determine for each recipe whether it appears as a contiguous subarray in the ingredient list.
Follow-up: can you solve this using only O(1) extra space?
Follow-up: now imagine ingredients arrive as a high-throughput data stream. How would you redesign the solution to handle each incoming ingredient efficiently?

“Started with a hash-prefix approach and talked through why it degrades to O(n^2) in the worst case when you enumerate substrings.”

View Post

Snowflake·Software Engineer·Technical Phone Screen

Apr 2026

Snowflake SWE interview with a grid-based assignment problem that builds on a classic cake-and-person setup. The core challenge was getting the greedy heap approach right and handling the follow-up constraint about determinism cleanly.

You have a 2D grid with people and cakes placed on it. Each person will eat the cake closest to them, and each cake can only be consumed by one person. Given a specific person's index, return which cake that person ends up with. You're guaranteed no two people share the same distance to any cake, and no two cakes share the same distance to any person.

“The constraints are what make this tractable.”

View Post

Snowflake·Software Engineer·Technical Phone Screen

Apr 2026

Snowflake software engineer interview with a classic scheduling DP problem. Nothing too surprising but the problem has enough moving parts that it's easy to fumble if you haven't seen interval scheduling before.

Given a list of jobs each with a start time, end time, and profit value, find the maximum profit you can earn by picking a non-overlapping subset of jobs. Jobs that share an endpoint (one ends exactly when another starts) are considered compatible.

“This is basically the weighted job scheduling problem.”

View Post

Snowflake·Software Engineer·Onsite - System Design / Architecture

Apr 2026

Snowflake system design round for a software engineer role. Started from a blank whiteboard, no scaffold, no hints. The problem looked manageable at first but the follow-ups kept stacking and by the distributed design portion I was definitely winging parts of it.

Design a service with two APIs: one to insert or replace a document by filename (tokenizing the text content into words), and one to check whether a document satisfies a boolean predicate over words, supporting AND, OR, and parentheses with standard operator precedence.
Extend the service with a GetAllFiles API that returns every document satisfying a given predicate. How would you design the supporting indexes to make this efficient?
How would you extend this into a distributed system? Discuss sharding strategies (by document vs by word), replication, query routing and fan-out, and what consistency guarantees you'd offer.

“The insert part was fine, just a map from filename to a set of tokens.”

View Post

Snowflake·Software Engineer·Technical Phone Screen

Apr 2026

Snowflake software engineer coding round, nothing too surprising. All questions were repeats from known prep sites, finished the first coding problem well under the time limit.

Given a game state, determine whether the current player can guarantee a win with optimal play (a 'can I win' style problem).

“Had seen this type before so I wasn't starting from scratch.”

View Post

Snowflake·Software Engineer·Technical Phone Screen

Apr 2026

Snowflake SWE interview with a graph traversal problem that looks simple on the surface but has enough follow-up layers to trip you up if you're not careful about the details.

You have a start URI and a target URI, and an API that returns all URIs linked from a given page. Find the minimum number of clicks to get from start to target.

“My first instinct was DFS and I had to stop myself.”

View Post

Snowflake·Software Engineer·Onsite - System Design / Architecture

Apr 2026

Snowflake software engineer interview, system design round focused entirely on building a key-value store from scratch. The problem escalated fast from basic CRUD to nested transactions to concurrency, which I was not fully ready for.

Design and implement an in-memory key-value store with get, put, and delete operations.
Extend the key-value store to support transactions, including nested transactions, with begin, commit, and rollback operations.
How does delete interact with transactions? If a key is deleted inside a transaction and then rolled back, what should happen?
What happens if commit or rollback is called when there is no active transaction?
How would you extend this design to support concurrent access from multiple threads? What consistency guarantees would you provide and how would you enforce them?

“Felt fine at first.”

View Post

Snowflake·Software Engineer·Technical Phone Screen

Apr 2026

Snowflake software engineer interview with a system design coding question about building an in-memory file system. Pretty classic OOP/design problem but there are enough edge cases to trip you up if you haven't thought it through.

Design an in-memory file system that supports ls (list files or directory contents), mkdir (create directories recursively), addContentToFile (create or append to a file), and readContentFromFile (return file contents as a string). All paths follow Unix-style absolute formatting.

“I started with a trie-like node structure where each node tracks whether it's a file or directory and stores children in a sorted map.”

View Post

Snowflake·Software Engineer·Technical Phone Screen

Apr 2026

Snowflake SWE coding round, one problem the whole session. Grid BFS thing that looks approachable until you realize they want optimal complexity and a clean multi-source setup, not just a per-desk Dijkstra loop.

Given a 2D character grid containing desks, bathrooms, empty spaces, and walls, find the shortest walking distance from each desk to its nearest bathroom using only 4-directional moves. Return a map of desk coordinates to minimum distances, or -1 if a desk has no path to any bathroom.

“My first instinct was to BFS from each desk separately and I even started coding it up.”

View Post

Snowflake·Software Engineer·Onsite - System Design / Architecture

Apr 2026

Snowflake system design round focused entirely on building a distributed job scheduler from scratch. The scope was pretty wide and covered everything from the core scheduling logic to HA, storage, and observability. A lot of ground to cover in one session.

Design a distributed job scheduler that supports one-shot and recurring (cron-style) jobs, with the ability to pause and resume jobs without killing in-flight executions.
How would you track job run history and status across states like queued, running, succeeded, and failed?
How do you ensure high availability for the scheduler itself, including leader election and failover behavior?
What are the trade-offs between exactly-once and at-least-once execution semantics for scheduled jobs, and what idempotency guarantees do you expect from workers?
How do you handle clock skew across nodes and missed cron fires after the scheduler experiences downtime?
How would you handle worker assignment and job dispatching given effectively unlimited worker capacity, and where does backpressure still apply?
What observability and rate limiting mechanisms would you build into this system?

“The pause/resume part is where I got a bit tangled.”

View Post