First round was with recruiter and is very easy and just a quick vibes/background check. Second round was also very easy, but in my opinion the success criteria is not clearly defined. The prep material directly states that AI use is allowed and we were expected to test critical business thinking and reasoning by looking at flight data. This flight data was split into CSV files like tickets, flights, etc. However, during my interview, I was not allowed to use AI because "the data could have sensitive data" and I needed to use Excel or some other tool. This was a curveball and I just quickly pulled up a collab session with pandas to write some code. The questions I asked were really basic like count how many nonstop flights exist in the dataset or find the cheapest economy flight. The most complicated code I wrote was a simple join statement on two dataframes. It's possible I was overthinking it from the start, but the framing behind the process strongly pushed to an AI first approach and diving deep. I remember experimenting with the test data before the interview to find data quality issues and inefficient flight hubs quickly with AI. The questions I was asked seemed misaligned with a Palantir funded startup that markets itself as AI first. I'm unsure of what was expected as I didn't get another round, but the process felt arbitrary and the success criteria is not clearly defined. I was told that the later rounds would have been a system design round and then there would be a founder talk, but I'm unsure what this would look like.