I went through the loop for a Software Engineer II backend position at Airwallex in Singapore. In total there were four interview rounds, plus one extra HR conversation where we talked about which team I'd want to land in. The four rounds were a coding interview, an AI-assisted coding interview, a system design interview, and a final Bar Raiser.
The coding round gave me a dynamic programming problem: given a triangle of numbers, find the path with the largest sum. The follow-up pinned one particular cell in the triangle and asked for the best path that has to go through that cell. To prepare, I mostly reviewed the usual LeetCode patterns and practiced explaining my thinking out loud.
In the AI-assisted coding round, I had to build a Tic-Tac-Toe game and then use AI tools to grow it into a proper app with a frontend and a backend. Beyond just writing the code, what mattered was using the AI well while still actually understanding what it produced, iterating on it, debugging problems, and checking that the end result worked.
The system design question was to design something like Booking.com. I had studied a system design framework beforehand and it helped a lot, and I took the Ticketmaster design as a rough template for my answer. Be aware that the interviewer pushed deep into implementation specifics, for instance asking me to lay out the real SQL tables and the important columns of each one. So I'd prepare for more than the high-level architecture.
The Bar Raiser was behavioral questions plus deep dives into projects I'd worked on. My advice is to have two or three solid projects ready and expect many layers of follow-ups about your decisions, the trade-offs, the impact, and what you learned. Also spend a little time coming up with some good questions to ask the interviewer at the end.
Overall I thought the process was practical and well organized, and the questions felt like the kind of work a software engineer would really do.
As for preparation, what helped me most was strong fundamentals and a clear, structured way to approach each problem. I'd also say don't stop at surface-level prep: be able to talk through implementation details in system design, be able to read and debug code that an AI wrote, and be ready to discuss your past projects in depth.
Discussion
Loading comments…