Square · Primly Community

Square coding interview / online assessment, format and difficulty in 2026

qa_quinn · 4 replies

Went through this recently for a backend SWE position. Here's the actual format so you're not guessing.

Online assessment (OA): Two coding problems, 90 minutes, HackerRank. Problems were medium difficulty by Leetcode standards. One was graph-adjacent (finding shortest paths in a transaction network, essentially), the other was string parsing with some edge cases around currency formatting. No system design in the OA.

Phone screen coding: One problem, 45 minutes, CodePair (collaborative editor). The problem was sliding window / two-pointer territory. Medium. The interviewer was active, asked me to explain my thinking as I went, and pushed me to optimize after I had a working solution. Real-time communication matters here, silence is bad.

Onsite coding rounds: Two rounds. One was more algorithmic (trees, I think BFS), one was more "design this API and write the supporting code" -- so half design, half implementation. The second type is more interesting and more signal-y in my experience.

Overall difficulty: solid LC medium. I didn't see anything that felt like hard-tier, but the "API design + code" round can catch you off guard if you've only drilled pure algorithm problems.

Language support: Python, Java, Go, Kotlin, Scala. Java/Python are most comfortable for interviewers to read.

A few notes: Clean code matters more than at some companies. Naming things well, not leaving dead branches, some basic error handling. They're building payment infrastructure, sloppy code reads as risk. Time complexity questions always come at the end. Have your Big-O ready. I was not asked any DP problems across the full loop, though others have been. Don't skip it entirely.

Total interview timeline from OA invite to offer: about 5 weeks. The OA was the first step, then recruiter call, then phone screen, then onsite (virtual). No take-home in my loop.

4 replies

consultant_cam

The OA problems -- were they straightforward "submit and it passes test cases" or did they have tricky edge cases that failed you on hidden tests? Asking because HackerRank can feel fine and then tank your score on hidden inputs.

sec_sasha

Honest answer: I don't know what the hidden test coverage was. I passed to the next round so I assume OK. I tested currency edge cases (zero values, negative amounts in adjustments) which probably helped. Don't just write the happy path.

ml_mike

The API design + coding round is interesting, I've seen this format at a few fintech companies. It's genuinely testing product thinking + code together. Like, can you define a clean interface before you implement it? Worth spending some time designing toy payment APIs as practice.

content_cole

Good writeup. One thing to add: the CodePair environment lags sometimes. Type a little slower than you think you need to, the delay between keystrokes and render can mess with your flow if you're used to a fast local editor.