Dr. Hejazi's research interests combine modern causal inference and statistical machine learning focusing on the development of innovative methods for statistical inference within a model-agnostic framework. This results in novel techniques that are (1) robust, by leveraging data-adaptive and flexible estimation procedures; (2) efficient (i.e., attaining minimal variance), by using cutting-edge non-/semi-parametric efficiency theory; and (3) assumption-lean, by incorporating current scientific knowledge while respecting its many limitations. His approach to developing statistical methodology emphasizes a problem-first translational philosophy -- that inferential techniques should be tailored to specific issues arising in scientific collaborations so that these new methods speak most directly and clearly to the underlying questions. Dr. Hejazi is usually, but not exclusively, motivated by topics from distribution-free (nonparametric) inference, semiparametric-efficient inference, high-dimensional inference, (targeted) minimum loss-based estimation, biased sampling procedures (e.g., outcome-dependent two-phase designs, capture-recapture), and adaptive experimental design (e.g., adaptive group-sequential designs for finding optimal sequentially adaptive treatments). Most often, he studies these topics through the lens of particular causal inference problems (e.g., heterogeneous treatment effects, dose-response curve estimation, mediation analysis). Dr. Hejazi is also deeply interested in high-performance statistical computing and is a passionate advocate of open-source software for the promotion of transparency, reproducibility, and "data analytic hygiene" in the practice of applied statistics and statistical data science.
Dr. Hejazi's substantive scientific interests across the biomedical and public health sciences are diverse -- accordingly, over the years, he has contributed to scientific collaborations across a broad array of disciplines, including in molecular and environmental toxicology, large-scale interventional nutritional epidemiology, comparative (health) effectiveness research, studies using electronic health/medical records, and computational/high-dimensional biology. Most recently, he has been captivated by the rich statistical issues and pressing public health challenges often found in clinical trials or observational studies seeking to evaluate the efficacy of preventive vaccines or curative treatments of high-burden and/or neglected infectious diseases (HIV-1/AIDS, COVID-19, malaria) and in infectious disease epidemiology (especially in the design of optimal minimally invasive surveillance strategies).
"Far better an approximate answer to the right question, which is often vague, than the exact answer to the wrong question, which can always be made precise." --John Tukey
"Everyone is sure of this [that errors are normally distributed]...since the experimentalists believe that it is a mathematical theorem, and the mathematicians that it is an experimentally determined fact." --Henri Poincare