The Science Behind the Scroll: Donghui Li’s Path from Cornell Tech to TikTok

By Sarah Marquart

When a user discovers TikTok through an ad, a friend’s recommendation, or a chance scroll through the app store, it might feel like a simple moment. For Cornell Tech alumna Donghui Li, though, it’s a complex data problem hiding in plain sight — and solving it at global scale is what drives her.

Li is a senior data scientist at TikTok, where she develops and improves attribution and growth measurement systems that touch a massive global community. She works to answer some of the app’s most consequential questions: Which marketing channels drive user engagement? What causes them to join? And how does their engagement evolve over time? These insights help cross-functional teams interpret growth signals with greater clarity.

“A metric is never just a number,” Li explains. “Behind it are diverse user needs, product contexts, and measurement challenges. To interpret the data well, I need to understand not only the statistical signal, but also the product dynamics and market context behind them.”

Li’s path to this work began long before TikTok was a household name, or even available internationally. After graduating with a degree in statistics from Beijing Information Science and Technology University in 2016, she spent several years in online advertising and data science.

There, she developed an early interest in measurement — specifically, how to determine whether a campaign or product change had truly caused an outcome, rather than simply coinciding with one. That question, she found, sat at the intersection of rigorous statistics and real business consequence.

She wanted to go deeper and sought out a program that could take her there. Drawn by its combination of technical rigor, entrepreneurial environment, and proximity to New York’s tech ecosystem, she enrolled in Cornell Tech’s Master of Science in Operations Research and Information Engineering program in July 2023.

“Cornell Tech was a turning point for me,” she says. “Returning to school gave me space to step back from day-to-day execution, strengthen my foundations, and organize my experience more systematically.”

Beyond the curriculum, Li credits Cornell Tech’s Studio model and cross-disciplinary environment with broadening how she thinks about her role. “It helped me think more about users, products, business context, and real-world impact,” she says. “Today, I see data science not as a boundary, but as one of the tools I can use to solve meaningful problems.”

At TikTok, that expanded perspective has translated into some of her proudest work — including improving the platform’s user growth attribution framework to make it more accurate, interpretable, and useful for cross-functional teams.

Her contributions to this field have extended beyond TikTok itself. Li is an inventor on multiple patents related to attribution and measurement systems and recently had a paper on advanced attribution methodologies accepted at AdKDD 2026, a premier workshop on AI for advertising held in conjunction with the ACM SIGKDD Conference on Knowledge Discovery and Data Mining. “My Cornell Tech experience really empowered me to better bridge the gap between academic research and real-world industry applications,” she says.

Looking ahead, Li is energized by AI’s potential to reshape how data scientists work, moving from pure analysis toward product building and decision systems. “The most exciting part is not just the technology itself, but how it pushes us to rethink how we work and how we create value.”

For current Cornell Tech students, her advice reflects her own journey: don’t wait to feel ready. “The best way to learn is often to start, experiment, and improve along the way.” And beyond the technical skills, she says, don’t underestimate the community you’re building. “The trust built during school feels very different from many professional relationships — and for me, it has continued to matter long after graduation.”

Sarah Marquart is a freelance writer for Cornell Tech.