Turing Software Engineer interview experience

The first one was positioned as an easy warm-up, and the second was medium, meant to test whether I could really solve instead of just stumble into an answer.

Right after the coding, I hit a more unusual technical assessment: evaluating the outputs of a generative AI model. I was shown responses to specific queries and had to compare them for accuracy and relevance, then answer follow-up questions that depended on the model’s output. That part felt less like standard programming and more like judging reasoning quality and picking apart what was wrong or weak.

The overall difficulty felt