Design Before Modeling
Many downstream problems get easier when the study question, comparator, index date, and estimand are thoughtfully defined at the outset.
RWE Focus
The real-world evidence work I care about most combines scientific rigor with operational practicality. Good studies do not happen by accident. They depend on clear design choices, disciplined execution, strong collaboration, and a willingness to think carefully about what the evidence can and cannot support.
In practice, that means caring deeply about things like comparator strategy, time zero, confounding control, endpoint quality, sensitivity analyses, and the reproducibility of the underlying workflow. It also means building systems that make good decisions easier for teams to repeat in biopharma environments where evidence quality and execution both matter.
What Matters
Many downstream problems get easier when the study question, comparator, index date, and estimand are thoughtfully defined at the outset.
Strong evidence generation depends not only on statistical methods, but also on analytic systems that are durable, reviewable, and repeatable.
The best work usually comes from close partnership across science, product, programming, and domain teams rather than isolated technical excellence.
AI can help accelerate internal workflows and support better tools, but it should be used in ways that strengthen scientific quality rather than obscure it.