No, it’s not a dumb question. Almost everyone on codechef, transitioning from competitive programming to real-world software dev has this question, and from my experience, sadly the answer is Nowhere.
The roles in the industry are mainly from Backend, frontend, Android, Big data, or ML. In none of these roles, you will need real DSA knowledge of the kind of used in competitive coding.
In research-oriented roles (maybe top 1% programmers) at Google Research, Microsoft Research etc use DSA and other deep computer science concepts.
As your experience says, DSA is rarely used in frontend (flask and react). DSA is sometimes needed in Backend, Big data or ML roles, where there is a large amount of data that needs to be processed fast or your service is getting lakhs of queries of users which need to be processed with minimal latencies. But, even there, we mostly used libraries and never code DSA on our own.
I interned at Google and worked on backend dev. Although I cannot give a lot of details about the project (due to NDA), the kind of DSA that we used was a BFS on a graph, caching of results and multi-threading / parallel computing. And, I doubt anyone uses more complicated DSA than this in regular work.
Maybe you will see binary search replacing linear search in some places, converting a loop into a map, so that you can access things in O(1), but not much complicated. I doubt anyone would be using segment trees, BIT, DP or such things in normal software dev anywhere.
Also, just a personal advice, try to work for small product-based startups rather than these big service-based companies (like TCS, InfoSys, Wipro etc). Also, if it works out for you, I would highly recommend doing an MTech from a tier 1 college (IIT/NIT/BITS etc) or if you can afford a foreign MS.