Lars van der Laan
University of Washington, Seattle. Department of Statistics

B313
Padelford Hall, Northeast Stevens Way
Seattle, WA 98103
lvdlaan@uw.edu
About me
I am a fifth-year Ph.D. student in Statistics at the University of Washington, advised by Marco Carone and Alex Luedtke.
My research focuses on causal inference, statistical learning, and semiparametric efficiency theory. I develop methods for debiased and efficient estimation with modern machine learning, including doubly robust inference, automatic debiasing, inference after model selection, and heterogeneous treatment effects.
I am supported by a Netflix Graduate Research Fellowship, collaborating with Nathan Kallus and Aurélien Bibaut on reinforcement learning and dynamic decision making (paper) and nonparametric instrumental variables inference (paper). This work tackles challenges such as estimating long-term causal effects from short-term experiments.
Another line of research connects calibration with causal inference, including causal isotonic calibration for CATE predictors, stabilized weighting, Bellman calibration for reinforcement learning, and doubly robust inference via calibration.
I also study machine learning methods for predictive inference, such as conformal prediction and generalized Venn–Abers calibration (with Ahmed Alaa, UC Berkeley).
Beyond methodology, I apply these ideas in biomedical and technology domains through internships at Genentech, the Fred Hutchinson Cancer Center, and Netflix. I also contribute to the tlverse open-source software ecosystem and consult for TLRevolution.