Canva Data Scientist Interview Guide

I got into Canva’s process for a Data Scientist role expecting a pretty standard mix of SQL and ML, but the technical depth was a little broader than I expected. The first step was a phone screen that felt mostly like a fit check — they were looking for whether my background matched their needs and values, and there was a straightforward “why Canva” question. After that came a one-hour advanced SQL round, which was very focused on practical querying rather than trivia. Window functions came up, and I was also asked to talk through the complexity of joins and functions, so they cared about both correctness and how I reasoned about performance.

The heavier round was the technical interview that ran about 3.5 to 4 hours and could be split across two days. That one covered Python, AI, system design, behavioral questions, and modeling. In my case, there wasn’t any LeetCode-style coding in the later rounds, but there was a HackerRank stage earlier with 9 questions total: 3 coding questions that were easy-to-medium and 6 multiple-choice questions covering stats, math, and general data science concepts. Another technical round asked me to implement clustering from scratch and write a basic evaluation script, which was more about understanding the algorithm than memorizing syntax. The final interview was ML system design around a classification problem with human intervention, plus behavioral questions. Overall it wasn’t especially hard, but it was broad, and they seemed to care a lot about fundamentals and how I explained tradeoffs. I didn’t get an offer, so my main takeaway is to be ready for SQL depth, basic stats, and at least one ML/system design discussion that goes beyond just model selection.