Reflection on my interview w. Figma

Last week I completed my OA for Figma's Summer 2027 SWE Internship. In this reflection I'll talk about my preparation, the actual assessment and how I went through that and what I'm taking away.

Getting ready was its own challenge. AI-assisted OAs don't have much precedent online, so I went in figuring it out as I went. Once I started I read through the instructions fully before touching anything and then fed those same instructions to the AI so we were working from the same brief.

From there I set my own ground rules: what I wanted built, how I wanted it structured, where things should live. I was specific about types, specific about output format, specific about what the tests needed to actually measure. Treat the AI as a collaborator, not a crutch.

During the assessment I set clear expectations upfront, stayed involved in reviewing the AI's output, and answered the guided questions as I went. What I didn't do well enough was manage my time. The 90 minutes felt like a lot (and then it wasn't). I ran short on the final question which was (unfortunately for me in the moment) a real self-critique of my own implementation. I also missed some edge cases around input validation, the kind of thing that's easy to skip when you're moving fast.

The feedback was honest and fair. My core implementation held up and the design write-ups were strong. But the gaps were specific: I missed some edge cases, some test coverage, and had an incomplete final question. Ultimately I wasn't selected to continue the process. While I'm not entirely pleased with the execution, I'm proud I got there and I plan to reapply next season.

Between now and then, a few things I want to focus on: drill edge case thinking, make input validation a reflex, and protect time for self-audit questions since those aren't bonus points, they're part of the problem. In the meantime, I'm back to building and adding to the portfolio.

Thank you to the Figma team for the opportunity.