Your solution to the anagram problem is very ingenious, indeed. I have seen this question a bunch of times and know all the "standard" approaches - but your solution, which i've never heard of before, is correct, clever and far superior. Kudos to you for coming up with it under such stressful conditions!
Your description of the interview process is very sad. The interviewers' attitudes and response induced an icky feeling in me.
I think that the attitude you encountered at Kaggle is an artifact of rather poor quality, insecure and insufficiently experienced / accomplished engineers. Whenever i interact with really accomplished and senior technical people, the conversations tend to be of stellar quality - they like to discuss actual past work in depth, are able to comprehend new and (perhaps unusual to them) concepts, and are generally a lot more respectful and pleasant. On the other hand, everything you described reeks of junior, inexperienced and rather insecure technical talent.
This reminds me of my own very-first tech interview. I had spent 5 years in a research lab, working on autonomous vehicles, writing software for things like signal processing algorithms, error correction codes, ML / reinforcement learning algorithms optimised to run on power-constrained devices and the like.
I went to interview with a VC-funded startup. They were looking for "experts in signal processing and machine learning experts". I figured my years of signal processing / ML experience might be a good fit. After the initial pleasantries, i get asked "How will you design a server that can detect anagrams of any word in the english language?"
For the life of me, i couldn't come up with a reasonable solution. I went home and hung my head in shame.