Duolingo Software Engineer interview typically runs 6 rounds: recruiter screen, technical screen, online assessment, pair programming, code review, technical design, and hiring manager chat. The process usually takes a few weeks and is fast-moving, with tight feedback and varied interviewer engagement.
$120K
Avg. Base Comp
$201K
Avg. Total Comp
6-8
Typical Rounds
2-4 weeks
Process Length
We’ve seen Duolingo lean hard on correctness under pressure, not just whether a candidate can sketch the right approach. Multiple candidates reported that passing the visible solution wasn’t enough: one was stopped by hidden test cases on CodeSignal, while another described a process that felt unforgiving even after a solid technical screen. That pattern tells us Duolingo is looking for engineers who can reason through edge cases, not just get to a working answer quickly.
A recurring theme is the company’s interest in how candidates handle code in realistic settings. One candidate was asked to review a pull request and call out code quality issues, while another faced a pair-programming demo with a senior engineer. Those experiences suggest Duolingo values people who can read unfamiliar code, spot flaws, and collaborate without needing the interview to be a one-way performance. We’ve also seen that the technical bar isn’t limited to algorithms; design came up as the hardest part for one candidate, which hints that they care about whether you can make tradeoffs and explain them clearly.
The non-obvious signal here is the interviewers’ style. Several candidates noted quiet or disengaged interviewers, and the feedback loop could be very fast. In practice, that means candidates need to stay steady even when the room feels cold or the prompts feel narrow. At Duolingo, the strongest signal seems to be a combination of clean execution and the ability to defend decisions when the interviewer pushes on details.
Synthetized from 2 candidates reports by our editorial team.
Had an interview recently?
Share your experience. Unlock the full guide.
Real interview reports from people who went through the Duolingo process.
The weirdest part of my Duolingo process was that it never really felt like a normal “solve the problem and move on” interview. I had a recruiter call first, then a technical screen where I got a standard preorder tree traversal question, plus some follow-up discussion about how to make the algorithm more optimal. That screen was pretty straightforward, but the feedback loop felt tight and a little unforgiving. I also had an online assessment on CodeSignal with four questions, and a lot of it leaned on graph traversal. I passed the visible parts, but I didn’t get all of the hidden test cases on one of them, which seemed to be enough to stop me from moving forward.
The onsite sounded more interesting than the screen. One round was a code review on CodeSignal where I had to review a pull request and point out code quality issues and errors, which was unusual compared with the usual LeetCode-style setup. There was also a whiteboard medium-level LeetCode problem, and a pair programming round with a senior engineer on a toy demo using VS Code codesharing. That last round felt more collaborative and they did ask some prep questions and about my job responsibilities, so it wasn’t purely algorithmic. Overall the process seemed to care a lot about correctness and hidden edge cases, and if your code doesn’t pass everything cleanly, that can be the end of it pretty quickly. I ended up not getting an offer, and the whole thing gave off pretty cold vibes from start to finish.
Prep tip from this candidate
Brush up on graph traversal and tree traversal, and make sure your solutions handle hidden edge cases cleanly since failing hidden tests seemed to matter a lot. Also be ready for a code review round where you critique a pull request, not just write code from scratch.
Share your own interview experience to unlock all reports, or subscribe for full access.
Sourced from candidate reports and verified by our team.
Topics based on recent interview experiences.
Featured question at Duolingo
How would you build an algorithm to measure how difficult a piece of text is to read for a non-fluent speaker of a language.
| Question | |
|---|---|
| Employee Salaries | |
| 2nd Highest Salary | |
| Closest SAT Scores | |
| Empty Neighborhoods | |
| Merge Sorted Lists | |
| Subscription Overlap | |
| Top Three Salaries | |
| Raining in Seattle | |
| P-value to a Layman | |
| Weighted Keys | |
| Retailer Data Warehouse | |
| Sum to N | |
| Google Maps Improvement | |
| Top 3 Users | |
| Bagging vs Boosting | |
| Maximum Profit | |
| Hurdles In Data Projects | |
| Download Facts | |
| Permutation Palindrome | |
| Liked Pages | |
| User Experience Percentage | |
| Distance Traveled | |
| Delivery Estimate Model | |
| Find the First Non-Repeating Character in a String | |
| Last Transaction | |
| Third Purchase | |
| The Brackets Problem | |
| Daily Retention Summary | |
| Basic Regex |
Synthesized from candidate reports. Individual experiences may vary.
After applying online, a recruiter reaches out to walk through the role and interview process. This first conversation is mostly an overview of the loop and basic background alignment. A recruiter screen follows the intro call and serves as an initial filter before technical rounds. Candidates described it as straightforward and focused on moving into the assessment stages.
This round includes a coding problem such as preorder tree traversal or a LeetCode-style data structures question. Interviewers also ask follow-up questions about making the solution more optimal and may probe bonus questions. Candidates complete a CodeSignal or HackerRank assessment with multiple questions, often with a strong emphasis on graph traversal and correctness. Hidden test cases matter, and failing one can stop the process even if visible cases pass.
One onsite round is a collaborative pair-programming session, sometimes using VS Code codesharing, with a senior engineer. The discussion can include prep questions and questions about your job responsibilities, not just pure algorithmic coding. Another onsite round is a CodeSignal-based code review where you review a pull request and identify code quality issues and bugs. This is more about reading code critically than writing a solution from scratch.
Candidates also face a harder technical design round that was described as the most challenging part of the loop. The interview appears to test system thinking and tradeoff discussion rather than a fixed algorithmic prompt. The process ends with an ungraded hiring manager conversation. In one experience, no questions were asked, making it feel more like a final check-in than a formal interview.