AI/ML systems are powerful predictors, but they are not designed to answer counterfactual “what if” questions. My research combines flexible learning algorithms with causal inference methods to generate actionable evidence and develop tools for better decision-making.
I'm a PhD candidate in Epidemiology at Harvard, working with the Division of Pharmacoepidemiology at Brigham and Women's Hospital / Harvard Medical School, YLab, and CAUSALab.
Before graduate school, I spent five years as a research scientist at a healthcare analytics startup, building tools for real-world evidence generation.
ABOUT
Source: @EikoFried on X
Causality-aware AI/ML for medicine and science
Highlighted publications
Adaptive prespecification framework for observational fesearch (Epidemiology)
Scalable confounding adjustment in high-throughput evidence systems (JAMIA)
*Rothman Prize, runner-up (for best paper 2025)
Durability of the single-dose Ad26.COV2.S vaccine (JAMA Open)
Research
I'm a PhD candidate in Epidemiology at Harvard, collaborating with the Division of Pharmacoepidemiology at Brigham and Women's Hospital and Harvard Medical School, YLab, and CAUSALab.
Previously, I spent five years as a research scientist at a healthcare analytics startup, building tools for real‑world evidence generation.
About me
Highlighted publications
Comparative host-coronavirus protein interaction networks reveal pan-viral disease mechanisms (Science)
DT-Transformer: A foundation model for disease trajectory prediction (arXiv)
Scalable confounding adjustment in high-throughput evidence systems (JAMIA)
Illustrating an adaptive prespecification framework for observational research (Epidemiology)
*Rothman Prize, runner-up (for best paper 2025)
Tabular foundation models as nuisance learners for causal effect estimation (arXiv)
Highlighted publications
Framework for adaptive pre-specification (Epidemiology)
Scalable confounding adjustment in high-throughput evidence systems (JAMIA)
Integrates population‑level and biological data to identify novel drug‑repurposing candidates
Evaluation of semi‑automated pipelines for high‑ dimensional proxy adjustment, highlighting robust hybrid outcome‑adaptive LASSO strategies
Illustrates adaptive framework for pre-specifying analytic flexibility with target trial emulations in COVID-19
Foundation model for longitudinal disease prediction trained on 57.1M EHR records from 1.7M+ patients
My research focuses on developing, evaluating, and applying causal inference methods to problems in biomedicine and public health. This work spans methodological development, causality‑aware AI, and applied studies of clinical interventions.
Applied Work
Methods
Causality-aware AI
Studies of medication effectiveness and safety using large‑scale data, with a focus on drug repurposing and comparative effectiveness
Combining causal inference and AI/ML for scientific discovery, precision medicine, and learning health systems
Causal methods and other tools that make trustworthy evidence generation more scalable