Yujun Zhou

I’m a data scientist with a PhD in Applied Economics. I spent several years on recommendation quality and integrity measurement at a large short-form video platform, and I now lead data science for enterprise AI systems at an advertising and media group. Across both, my job is to define an outcome worth moving, test whether the evidence supports the decision, and build whatever is missing to check it.

The through line is measurement under pressure. Engagement moves for reasons that have nothing to do with a better product. An assistant that reads well can still fail the task it was hired for. Most of my work is deciding which comparison is credible enough to act on, and saying plainly when it isn’t.

I build the systems I need to test my own ideas — evaluation pipelines, data contracts, and small products with real users — because a claim I cannot reproduce is not evidence I can defend.