I agree that is hard to remember every little detail for every project you have worked on, but I can definitely recollect how I approached the biggest problems on almost every project. Usually these are the ones worth talking about.
The joys of age plus extremely-poor autobiographical memory.
It doesn’t help that what I think of as actually hard, day to day, is shit like documentation from FAANG companies lying to me and wasting a whole fucking day of my time, terrible error messages, mismanagement of upstream projects, bureaucracy making a two-day task take four weeks, and, in some common ecosystems, awful design of core tools and major libraries, or horribly wasteful library churn. “Hard problems” are nothing next to bullshit problems.
Which is to say that usually the hard stuff is an accidental request that nobody actually wants to pay for, so the only thing “hard” about it is recognizing that the client/stakeholder has wandered into territory they really ought not.
That said, I think that someone could remember a few things for past projects when preparing for an interview. Of course doing that on the spot might be very hard.