Something getting lost in the responses here is the "stress" factor of these interviews. When encountering a difficult architectural or algorithm decision, experience has taught me to develop slowly, prototype often, investigate previous work, and even step away from the issue. Meanwhile, engineering interviews require you to code immediately, in isolation, and then be judged on the first solution you crap out of your brain.

In a recent interview, my solution to a question was bad, I knew it was bad, and I informed my interviewees it was bad. It was difficult for me to come up with other solutions on the spot under pressure of judging eyes and limited time. Immediately after getting off the phone I was able to assess better ways to approach and solve the problem, but too late, those 45 minutes were more important to them than my many years of provable, directly related experience.