Published

23 June 2025

I am an assistant professor of biostatistics at the Harvard Chan School of Public Health, where I lead and organize the NSH Lab (pronounced like “niche”)—a (bio)statistical science research group that is focused on developing novel theory, methods, algorithms, and open-source software tools for causal inference and causal (or debiased or targeted) machine learning, non-parametric statistics, statistical machine learning, model-agnostic inference, and applied semi-parametric theory for causal machine learning. My statistical methods research is motivated by data-driven, real-world questions that arise from collaborations with applied biomedical and public health scientists. Specific areas of focus have included infectious disease sciences (especially the evaluation of investigational therapeutic agents and immune correlates of vaccine protection), the study of chronic diseases (e.g., Long COVID), and cancer science—and comparative effectiveness research and implementation science in these areas.

Prior to joining the faculty in the Department of Biostatistics at the Harvard Chan School of Public Health in 2022, I held an NSF Mathematical Sciences Postdoctoral Research Fellowship, sponsored jointly by Iván Díaz and Peter Gilbert, during which I developed new techniques for causal mediation analysis while serving as a core member of the COVID-19 Prevention Network’s biostatistics response team, to which I contributed statistical methods innovations and the design and implementation of open-source software tools for the evaluation of candidate immune correlates of vaccine protection. In 2021, I obtained my PhD in biostatistics from UC Berkeley, under the joint supervision of Mark van der Laan and Alan Hubbard, for work on statistical and causal inference approaches to assessing the impacts of stochastic treatment regimes. I also co-founded and served as a core developer for the TLverse project, an open-source software ecosystem for targeted machine learning. From 2011 to 2015, I happily spent time as an undergraduate—also at UC Berkeley—in beautiful Berkeley, CA.

John Tukey once famously noted that “the best thing about being a statistician is that you get to play in everyone’s backyard.” I think he got it right. My past collaborations have spanned diverse areas of the biomedical and public health sciences—from molecular toxicology and computational biology to environmental health and nutritional epidemiology. In the last few years, the rich statistical science problems that abound in the infectious disease sciences, especially in efforts to study investigational therapeutics and preventives, have captured my attention. My work in these areas has contributed novel methods and insights for identifying immune correlates of protection (or surrogate endpoints) in vaccine efficacy trials of HIV and COVID-19; for comparing the efficacy of therapeutics for MDR-TB in key populations and synthesizing evidence to support tailoring normative treatment guidance for people living with HIV; and for characterizing and shedding further light on Long COVID.

Despite well-meaning recommendations against doing so, I continue to try to write code for statistics research…collateral damage from this ill-advised pastime is logged on GitHub and visualized below.

Nima Hejazi's GitHub Activity


In my free time, I enjoy trail running and training for half-marathons, experiencing live music, and trying my hand at—amateur, for now—landscape and urban photography.

Finally, here are a few reflections on science and statistics, mostly ones I wish I’d come up with first.

“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

“Although this may seem a paradox, all exact science is dominated by the idea of approximation.” –Bertrand Russell

“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

“The anarchy of guess and intuition has given way to a benevolent tyranny of statisticians.” –Donald Fredrickson

Nima Hejazi is an assistant professor of biostatistics at the Harvard T.H. Chan School of Public Health. His research program explores how causal inference, statistical machine learning, and computational statistics catalyze discovery in the biomedical and public health sciences. His work focuses on the development of model-agnostic, assumption-lean statistical inference procedures, emphasizing a translational philosophy that stresses the rich interplay between the applied sciences and statistical methods development. To do so, his approach draws upon causal inference principles to translate scientific questions into interpretable statistical estimands, which one can then aim to learn from data by designing analytic methods that (1) avoid imposing restrictions not justified by domain knowledge; (2) incorporate flexible, adaptive modeling strategies (i.e., machine learning); and (3) apply semi-parametric efficiency theory to achieve best-in-class uncertainty quantification. He is also interested in open-source software and high-performance computing for statistical science—both to push the boundaries of statistical methodology and to promote transparency and reproducibility in applied statistics.

His methodological work draws upon tools and ideas from semi-parametric statistics, high-dimensional and large-scale inference, de-biased and/or targeted machine learning (targeted minimum loss estimation, sieve estimation), and computational statistics. Recent areas of focus have included analyzing data collected using two-phase, auxiliary- or outcome-dependent sampling designs; evaluating causal effect heterogeneity and deriving optimal treatment rules to implement precision medicine; quantifying dose-response phenomena and the causal effects of continuous treatments; investigating questions of mechanism via causal mediation analysis; credibly drawing causal inferences when data are subject to interference or network dependence; and synthesizing evidence across both randomized trials and observational studies via data fusion.

John Tukey once noted that “the best thing about being a statistician is that you get to play in everyone’s backyard.” He got it right. Nima’s past collaborations have spanned diverse areas of the biomedical and public health sciences—from toxicology and computational biology to environmental health and nutritional epidemiology. More recently, though, he has been captivated by the rich statistical science problems that abound in the infectious disease sciences, especially in efforts to study investigational therapeutics and preventives, both in and across randomized controlled trials and observational studies. His work has contributed novel methods and insights for identifying immune correlates of protection (or surrogate endpoints) in vaccine efficacy trials of HIV and COVID-19, for comparing therapeutic efficacy in studies of COVID-19 and TB/HIV co-infection, and for characterizing and shedding further light on Long COVID.

Beginning in 2022, on numerous occasions—most often in Europe but at times in the US too—I’ve been stopped and politely informed that I bear a close resemblance to the Spanish footballer, Marc Cucurella. While I enjoy (watching) football and find the attention amusing, alas, I do not share Mr. Cucurella’s skills.

Can you spot the difference?

Left: A statistician who often wishes he were most anything else—why not a footballer?
Right: A footballer who has almost surely never wished he were a statistician.