← Back to Directory

Snap interview notes, page 12 of 22

Large Enterprises

Most of these Snap write-ups come from Software Engineers (117), then Machine Learning Engineers (65) and Data Scientists (18). The most reported rounds are the technical phone screen (149), the system design round (45), the behavioral / leadership round (11) and the product sense round (8).

216 interview notes · updated Jul 2026

Snap·Machine Learning Engineer·Onsite - System Design / Architecture

May 2026
Had a system design round at Snap for an ML engineer role, and the whole thing basically came down to one meaty question about RAG chunking strategy. Not what I expected to spend 45 minutes on, but here we are.
  • How would you design an effective chunking strategy for information in a RAG system?

“I went straight into fixed-size vs semantic chunking and talked about overlap windows, which felt solid at first.”

View Post

Snap·Machine Learning Engineer·Technical Phone Screen

May 2026
ML engineer loop at Snapchat, pretty heavy on fundamentals and deep learning theory. Five questions total and they did not let up on the math side of things. Left feeling like I probably over-explained some parts and under-explained others.
  • Walk through a project where you fine-tuned a large language model or foundation model, covering the task, how you built and labeled the dataset, whether you used full fine-tuning or a parameter-efficient method, your loss function, evaluation metrics, deployment constraints, and how you'd handle overfitting or hallucinations.
  • Explain regularization and compare L1, L2 and weight decay, dropout, early stopping, and data augmentation.
  • Compare SGD, SGD with momentum, Adam, and AdamW as optimizers, and explain when you'd pick each one.
  • Explain how self-attention works and extend that to multi-head attention, including the key equations.
  • Describe the full Transformer architecture and write out the main mathematical steps inside a decoder-style Transformer block.

“This was the one I actually felt okay about because I had a real project to pull from.”

View Post

Snap·Product Manager·Onsite - Product Sense / Strategy

May 2026
PM interview at Snap that threw a classic estimation question at me. Nothing too wild but it required some on-the-spot structure.
  • How many windows do you estimate exist in Tehran?

“A Fermi estimation question with a regional twist.”

View Post

Snap·Backend Engineer·Technical Phone Screen

May 2026
SQL-heavy technical screen for a backend role at Snapchat. The interviewer gave me a multi-table schema upfront and then fired off a few verbal requests one by one, expecting me to write queries on the spot and talk through the logic without actually running anything.
  • Given a schema with users, files, groups, and ACL tables from a file-access system, write a SQL query that joins two or more of these tables to answer a specific access-related question, and walk through your reasoning without executing it.
  • Write a GROUP BY query with aggregation across multiple tables from the same schema, and explain how you'd handle deduplication if a join produces duplicate rows.
  • For a given verbal data request, explain how your SQL query maps back to the underlying data model and why the join type you chose is correct for that specific access pattern.

“The part that tripped me up was deciding between INNER and LEFT JOIN before I fully understood what NULLs in the ACL table would mean for the result.”

View Post

Snap·Software Engineer·Technical Phone Screen

May 2026
Snapchat software engineer interview focused on mobile state management, specifically building out a color-picker feature across views and defending your architectural choices. The technical depth caught me a bit off guard since it went well beyond just coding a solution.
  • You have a scrollable grid where tapping a cell changes its color and pushes a detail view showing that color. Extend this so the user can modify the color on the detail screen, and when they navigate back, the grid cell reflects whatever they changed it to. How do you implement this?
  • Walk through the different patterns for propagating state changes back from a child view to a parent in SwiftUI versus UIKit, and explain the trade-offs.

“The coding part was fine, the follow-up was where it got interesting.”

View Post

Snap·Software Engineer·Onsite - System Design / Architecture

May 2026
System design round at Snapchat for a software engineering role. The whole thing was a deep crawling and catalog infrastructure problem, way more involved than I expected for a single session.
  • Design a product catalog system for an ad company that crawls third-party e-commerce sites to build its own structured catalog, without any data feeds from the merchants.
  • How would you handle crawler politeness, specifically robots.txt compliance and rate limiting across many partner sites?
  • Per-site HTML parsing templates versus ML-based extraction: what are the trade-offs and when would you use each?
  • How would you deduplicate products that appear on multiple partner sites, given no shared SKU standard?
  • How would you optimize storage costs for the media assets (images and videos) you're pulling from partner sites?
  • What are the legal and terms-of-service risks involved in crawling partner sites, and how would you mitigate them?

“This is a beast of a question.”

View Post

Snap·Software Engineer·Technical Phone Screen

May 2026
Snap SWE interview with a graph/ratio problem that sounds deceptively straightforward until you actually try to implement it cleanly.
  • Given a list of conversion relationships between symbols (e.g. currency or unit pairs), write an algorithm that returns the conversion ratio for any given pair of symbols.

“This is basically a graph traversal problem in disguise.”

View Post

Snap·Software Engineer·Onsite - System Design / Architecture

May 2026
Snapchat system design round, one big question that sprawled into a bunch of sub-problems. The redirect path vs. counter latency tradeoff is where things got interesting and where I probably lost some points.
  • Design a URL shortening service that also tracks click analytics, including total clicks, clicks over time, and clicks by region.
  • How would you handle the counter increment on each redirect without adding latency to the redirect path itself?
  • Walk through your approach to short-code generation. What are the tradeoffs between using a distributed counter with base62 encoding versus a hashing approach?
  • How would you use caching or a CDN to reduce latency for redirects at a global scale?

“I started with short-code generation and spent probably too long debating counter-based base62 vs hashing.”

View Post

Snap·Software Engineer·Technical Phone Screen

May 2026
Snapchat software engineer interview with a workflow improvement question that sounds straightforward until you realize they want specifics on measurement and technical blockers, not just a vague 'we adopted CI/CD' story.
  • Tell me about a project where you improved your team's development workflow. What changes did you make, why were they necessary, what technical obstacles came up, and how did you actually measure whether it worked?

“This one tripped me up more than I expected.”

View Post

Snap·Machine Learning Engineer·Technical Phone Screen

May 2026
Snapchat MLE interview with a graph problem that looks straightforward until you realize they want you to actually know your cycle detection cold. Nothing too wild but the expected complexity constraint made it clear they weren't looking for a brute force attempt.
  • Given a number of courses and a list of prerequisite pairs, determine whether it's possible to complete all courses without getting stuck in a circular dependency.

“Classic topological sort / cycle detection disguised as a scheduling problem.”

View Post