Congrats Spring 2025 SDS Graduates!
Please join us in congratulating our Spring 2025 SDS Graduates! Postdoctoral scholar Tijana Zrnic and PhD scholars Radhika Kapoor and Justin Young reflect on the transformative impact of their time in the program. While their stories are just a few among many, these three individuals are especially notable examples of the talent and dedication found across the SDS scholar and fellow community. Their journeys capture the essence of what makes SDS unique: a dynamic, interdisciplinary environment that fosters both technical excellence and deep intellectual curiosity.
Tjiana Zrnic
I loved our weekly SDS group meetings. They made me feel integrated with the broader campus community, far beyond my department, and I got to learn about the best work from all schools on campus! This is a very special experience that a typical PhD/postdoc program does not offer.
SDS Impact on Growth
Over the course of the program, Tijana gained a deep appreciation for working across disciplinary boundaries. She learned extensively about areas far beyond her own and came to value the unique perspectives that emerge from bringing together an interdisciplinary group. Regular “Data & Donuts” discussions offered an opportunity to step back and consider broader questions in data science, academia, and beyond—an experience that stands in contrast to the often narrow focus of a traditional PhD.
Favorite Project
Among her favorite projects was active statistical inference, presented at ICML 2024. This method leverages AI to assist in data collection and inference while rigorously preserving the validity of results, even when the AI systems involved are biased. Its versatility has made it valuable across a range of data science applications. A follow-up collaboration extended the method into computational social science, resulting in a user-friendly tutorial that integrates the technique with widely used platforms like Prolific and cutting-edge large language models.
Looking Ahead
Starting in January 2026, Tijana will hold a joint appointment at Stanford, in Statistics and Management Science & Engineering, with a courtesy appointment in Computer Science.
Radhika Kapoor
My favorite feature [of the SDS Program] might be the SDS Slack channel, where you can ask almost any question and be sure to get a response – and unlike AI, the response is almost guaranteed to be correct and from a human.
SDS Impact on Growth
One of the most meaningful parts of Radhika’s journey through the Stanford Data Science (SDS) program wasn’t a single project or academic milestone—it was the extraordinary support she received from the SDS community during a deeply personal time. As she navigated her pregnancy and prepared for childbirth, the community's warmth and encouragement stood out. Sharing a personal life update in a professional setting felt vulnerable, but the response from the SDS team made a lasting impact. The leadership's enthusiasm, the excitement from fellow SDS scholars, and even a baby shower organized by the SDS team were all part of a series of moments that made Radhika feel genuinely seen and supported. In the final weeks of pregnancy, the SDS scholar space became a particularly comforting and safe place to work.
Favorite Project
Throughout her time in the program, Radhika found tremendous value in the SDS scholar community and in the guidance of senior faculty like Naras (Balasubramanian Narasimhan) and John Chambers. The interdisciplinary nature of the program broadened her perspective on how data science is applied across fields. She especially appreciated the informal, collaborative spirit, highlighted by the ever-active SDS Slack channel, where questions (ranging from technical to logistical) are met with helpful, human responses. “Unlike AI,” she joked, “the response is almost guaranteed to be correct and from a human!”
Academically, Radhika focused on using psychometrics and data science to improve how educational outcomes, such as math and reading performance, are measured in international contexts. Her work explores how data from international assessments can inform policy and facilitate meaningful cross-country comparisons. She also investigates the potential of large language models (LLMs) to generate high-quality test items, with the broader goal of supporting the development of robust data for global education.
Looking Ahead
Radhika will join Amazon as a Research Scientist at the end of the summer. As she closes this chapter, she leaves behind a legacy of scholarly excellence, community connection, and personal courage—hallmarks of the SDS experience.
Justin Young
SDS Impact on Growth
For Justin, the Stanford Data Science (SDS) program has been as much about meaningful personal connections as it has been about academic discovery. Reflecting on his time in SDS, he highlights not one defining memory, but a tapestry of friendships and moments shared across two transformative years. From bonding with Emmanuel Balogun in their first year over a shared love of art and staying active, to reconnecting with Léon in Paris during the winter break, the relationships forged through the SDS scholar community have left a lasting impression. Justin notes with gratitude how these friendships likely wouldn’t have happened without SDS.
Favorite Project
Academically, SDS has significantly broadened Justin’s perspective. Weekly scholar meetings, conferences, and seminars revealed just how deeply AI is transforming a wide range of disciplines. As an econometrician, this interdisciplinary exposure offered valuable insight into the challenges others face in applying data science and helped Justin see where new tools and approaches might be needed most.
At Stanford, Justin's research has focused on integrating machine learning into the estimation of treatment effects—a central concern for anyone working in policy evaluation or applied statistics. One of his recent projects, in collaboration with Microsoft Research, has tackled the credibility of observational methods for causal inference. Drawing on feedback from across the SDS community, including scholars working on causal inference in disciplines beyond economics, Justin and his collaborators were able to sharpen the project’s direction and impact.
Why It Matters
Many applied researchers, from policymakers to data scientists in tech, often run experiments to determine the effect of an intervention, but in the absence of an experiment, they often have to rely on observational methods, whose credibility is always limited. Many advancements have been made in the econometrics literature to put forth best practices in order to recover the unseen ground truth. Justin's research extends this line of work into modern, high-dimensional datasets and incorporates all the recent developments in causal machine learning methodology that have made it easier to estimate flexible relationships between confounders, treatments, and outcomes. Justin and his team have introduced a new data sample encompassing an experimental rollout of a new feature at a large technology company and a simultaneous sample of users who endogenously opted into the feature. They have found that it is possible to recover ground truth causal effects but only with careful choices in modeling, which we distill as guidance for practitioners.
Looking Ahead
Justin's interest in real-world data and causality led him to two internships at Microsoft Research, where he deepened his focus on causal inference using industry data. Justin is excited to continue this line of work as a researcher in Microsoft’s Office of the Chief Economist. Whether through research or relationships, Justin leaves SDS having made lasting contributions and connections.
