I'll begin by setting the two mentioned cases aside. In the case of Rohit Vemula, the allegations about caste based discrimination aren't very strong. The two points are about the denial of his stipend and his suspension from the college. The college authority says that the stipend was delayed due to paperwork issues, and he was suspended following allegations of assaulting an ABVP member. Both explanations seem totally plausible. And, in the case of Payal, the harassment that is often talked about is "cultural", as I discussed earlier.

Usage of casteist slurs, etc is cultural. Please show me data about exclusion from hostels/labs. Even if it exists, it is already illegal in India and proper action should be taken to stop this. Also, I am interested to know more about the evidence for caste biased grading. The reason for higher proportion of dropouts is slightly complicated. I'll explain for IITs because that is what I'm most familiar with.

Two years back, friend of mine hacked in the academic website of IITM and stole the CGPA details of all the students. A few people did some interesting analyses, for example, if you sort the CGPAs, the bottom 10% in any batch were all scheduled groups. This shouldn't be surprising when the reservation system of India is so liberal. Take for example, Electrical Engineering branch at IITM. For general category you'd need a rank of around 500-1000; for SC Male, this should be around 12K. It should be even lower for SC Females and STs. Now, to get a rank of 500-1000 you'd need around 200 marks in JEE Advanced (I know about 2022, cuz that's when I attempted). And, for 12K, you'd need around 80 marks. The difference in the cognitive and other relevant abilities of these students is so huge that the SC Male can barely survive in a system that grades on relative scale. No wonder most people who get U grades (fail at IITM) are often from Scheduled groups. This observation is also substantiated from Placement data. Most people who are not placed in phase 1 are from scheduled groups.

If this is how much these students are struggling with the academics, no wonder they are represented highest among the dropouts, but people who lack the situational awareness immediately reason caste based structural discrimination.

I am unsure how much surname impacts interview selections but I won't be surprised if there turns out to be a bias (implicit or intentional). Not to mention that it's genuinely quite hard to find why exactly a candidate is rejected in an interview. Anyway, I'm open to discuss potential solutions here. Again, for getting callbacks, I'm unsure. Won't be surprised again. I guess, it's mostly implicit just like female names get more callbacks.

Regarding reservations, as you said, the objective was to give the marginalised some opportunities, but that has failed so catastrophically on all metrics that it is laughable to talk about the leveling.

Coming back, I think most examples that you have laid out here are either misplaced or represent cultural casteism. We should try our best to reduce that but that is categorically different from institutional casteism, which is rightly illegal and also virtually inexistent.