Turing Software Engineer interview experience

The first two rounds pulled from LeetCode-style questions, and the feedback was limited: I couldn’t see the full test cases, only the percentage of tests that passed. That made progress feel a bit like solving in the dark, but it also created a strong sense of momentum—either the logic worked, or it didn’t.

The third round was different and more focused on LLM output evaluation. I was shown two outputs from LLM prompts, and I had to identify the correct one and then write a detailed explanation. That writing and justification step felt like a major shift from the pure coding