Went through the product designer loop at OpenAI earlier this year. This is not a typical big-tech product design interview and the portfolio review component specifically caught me off guard in a good way.
Let me be direct about what they care about: designing for fundamentally new interaction paradigms. Not making a settings page clearer, not optimizing a checkout funnel. They're thinking about how humans interact with AI systems in ways that don't have established patterns yet. If your portfolio is strong on polish but thin on novel problem-solving, that will show.
The process: recruiter screen, portfolio review call with a senior designer (45 min), then a full design onsite: a design exercise, a portfolio deep-dive, a cross-functional round, and values.
Portfolio review: They're going to push on your process more than your output. Pick 2-3 projects that show how you navigate ambiguity, how you think about constraints, and where you made hard tradeoffs. The best conversation I had was about a project that didn't go perfectly. They asked what I'd do differently. Know your case studies deeply.
Design exercise: I got a prompt about an AI-adjacent product problem. You have 30-45 minutes then present. There's no objectively correct answer. They want to see how you frame the problem, what questions you ask, how you think about user mental models when the underlying technology is non-deterministic. Sketch quickly, explain your reasoning, show you can make decisions under uncertainty.
Cross-functional round: Basically: how do you work with PMs and engineers when the product direction isn't clear. Very collaboration-focused. Have examples.
Values round: Thoughtful conversation about your relationship to AI tools as a designer, what responsible AI product decisions look like, how you think about users who don't understand what the system is doing.
One thing: they moved fast in my process. Recruiter was responsive, decisions came within a week of each round. Rare.
Comp for senior product designer in SF: roughly comparable to Google/Meta at the same level, maybe slightly below on cash but with equity upside that could matter more. Depends heavily on leveling.