Alexander Sietsema
I will be on the job market beginning in September 2026.
I am a fourth year applied mathematics Ph.D. candidate at UCLA advised by Deanna Needell. My research focuses on better understanding the mathematical tools of data science and machine learning. As modern datasets continue to become larger and more complex, understanding their properties and how to learn from them has become increasingly important. Many datasets contain natural structures like symmetry, smoothness, or intrinsic low-dimensional geometry; simultaneously, the methods we choose to learn from these datasets induce properties of their own. My research seeks to better understand how such structures and inductive biases affect the performance of learning algorithms. I work to identify these structures, prove theoretical guarantees about their effects, and incorporate them into new, more effective learning methods. I approach these problems using tools from high-dimensional probability, numerical linear algebra, and computational harmonic analysis, working across both theory and practice. See the Research section for more details.
Before starting at UCLA, I received a B.S. in Advanced Mathematics and a B.S. in Computational Mathematics from Michigan State University. I also grew up in East Lansing. See the Personal section for more details.
Recent News
May 2026: I have been selected as one of the recipients of the UCLA Department of Mathematics Dissertation Year Award for 2026-2027! I am grateful to the department for their recognition and support.
May 2026: I have received the Liggett UCLA Math Department Teaching Award for 2026!
April 2026: Our paper “Harmful Overfitting in Sobolev Spaces” has been accepted to ICML 2026! I will be presenting in person in July at the conference in Seoul, South Korea.
April 2026: Our paper “VDW-GNNs: Vector diffusion wavelets for geometric graph neural networks” has been accepted to ICML 2026! I will be presenting in person in July at the conference in Seoul, South Korea.
April 2026: I will be participating in the University of British Columbia Women in Mathematics of Data Science Workshop! The workshop will take place from August 4-14 in Vancouver, BC, Canada.
April 2026: I will be speaking at the International Conference on Computational Harmonic Analysis at Vanderbilt University in May! The conference will be held from May 18-22 in Nashville, TN.
April 2026: I will be giving a talk at the Southern California Applied Mathematics Symposium at Loyola Marymount University in late April! I will be speaking about our recent work in harmful and benign overfitting in Sobolev spaces.
February 2026: I have been accepted into the ICTP-INdAM-SLMath summer school on Mathematics For Machine Learning! The summer school will be held in Trieste, Italy from June 15-June 22.