It was a one round interview. It was predominantly on data science and large language models, they gave me a repo of an agent framework, and asked me to find the code where reward is calculated.