Snap SWE Interview: Onsite Coding Rounds Guide

Estimated read time: 8-10 minutes

Summary: Snap SWE onsite coding rounds can include multiple engineer-led coding interviews, with one candidate report describing three coding rounds in a larger panel. Prepare for repeated Craft signal: clean implementation, edge cases, communication, and adaptation to follow-ups.

TL;DR + FAQ

• Candidate evidence supports multi-round onsite coding in some loops.
• Reported themes include graph conversion ratios, BFS/DFS, powers of two, product tradeoffs, and role-specific technical work.
• Round count varies by team and level.
• Craft matters across repeated rounds: clean code, tests, tradeoffs, and adaptation.
• Senior candidates should connect implementation to product or architecture when asked.

Quick FAQ

How many coding rounds are there?
Some reports mention several, but the source does not prove a universal count.

Are mobile or ML rounds different?
They can be. Treat those reports as role-specific.

Is product discussion part of coding?
It can be, especially when a task is feature-shaped.

What should I show repeatedly?
Clear reasoning, correct code, tests, complexity, and follow-up handling.

---

1) How onsite coding works

Onsite coding rounds give multiple engineers a chance to evaluate Craft. You may solve standalone algorithmic tasks, feature-shaped coding tasks, or role-specific implementations.

---

2) Onsite coding tasks you may face

• Given conversion ratios between symbols, answer conversion queries and handle missing links, cycles, and inconsistent data.
• Use BFS or DFS to search a graph, then change the task to return a path, shortest path, or all reachable nodes.
• Determine whether a value is a power of two, then extend the function to validate a stream of values efficiently.
• Build a feature helper that ranks or filters candidate items, then handle changing scores or product constraints.
• Implement a cache or session helper for a Snap feature and add expiration, invalidation, or device-specific behavior.
• For mobile roles, implement or explain a Swift/iOS component and handle concurrency, lifecycle, or data-refresh edge cases.
• For ads or ML roles, merge scored candidate lists and handle ties, missing scores, or diversity constraints.
• After the base solution works, explain how you would test it before launch.

---

3) What interviewers may probe

Expect probes on edge cases, test strategy, complexity, and product constraints. If the interviewer moves from generic code into a Snap feature, restate the product behavior before changing the implementation.

---

4) Common failure modes

Only solving the first version. Follow-ups can change the data structure.

Skipping tests. Multiple coding rounds make weak testing visible.

Ignoring product constraints. Snap feature context can affect correctness.

Overgeneralizing mobile reports. Mobile specificity should be labeled and confirmed.

---

5) How to prepare

• Practice graphs, BFS, DFS, maps, bit operations, caches, ranking, and streaming updates.
• Prepare role-specific tasks for mobile, AR, ads, ML, infra, or platform.
• Test every solution with normal, edge, invalid, and scale cases.
• Practice adapting when constraints change.
• Ask how many coding rounds are expected in your loop.

---