System design stopped being terrifying once I stopped trying to design the "perfect" system

Used to freeze up trying to land on the one correct architecture before saying anything out loud. What actually helped was starting with something obviously too simple, saying the assumption behind it, and then iterating from there in front of the interviewer instead of in my head. Turns out most of them want to see how you think through tradeoffs, not whether you memorized the textbook answer on the first try.