Green Buoy is the studio of Eric Green, Ph.D. — a quantitative scientist who designs research, models data, and builds AI products end to end.
AI assistants, consumer hardware, education tools, sports analytics, and a book — built solo or as co-founder, from research question to shipped product.
Prediction, causal inference, post-stratification, and NLP — the analytical backbone behind the products, published and reproducible.
Developed and validated a model linking self-reported readiness to subsequent healthcare-seeking — turning a simple stated-intent signal into a forward-looking prediction for targeting outreach and forecasting activation.
Used MRP to recover trustworthy population- and subgroup-level estimates of vaccine hesitancy from imperfect survey data — the technique that makes survey, panel, and real-world data decision-grade. Published as a reproducible tutorial.
Mined free-text conversations between consumers and a digital health service at scale to surface what people actually ask and which messaging resonates — the NLP foundation for voice-of-customer and conversational-AI analytics.
Designed a pilot randomized encouragement (instrumental-variable) trial to estimate the causal effect of a digital intervention when only the encouragement to use it can be randomized — the design logic behind real-world effectiveness studies.
Developed, validated, and shortened clinical screening instruments — from a perinatal depression tool that blends Western criteria with local idioms in Kenya, to individual-participant-data meta-analyses showing where PHQ-9 and EPDS scores do, and don’t, estimate true prevalence. The basis for trustworthy outcome measurement at scale.
Three threads, one practice: teaching rigorous methods, consulting on research and data for pharma and life-sciences, and building AI products that ship.
Causal inference, study design, and real-world evidence — questions framed so the answers hold up.
Modeling, inference, and analysis you can defend — with transparency and methodological rigor.
Pipelines, applied ML, and decision support that turn messy data into something teams can act on.
Turning rigorous analysis into decisions — clear insights and data storytelling for pharma and life-sciences teams.
LLM-powered products designed and built end to end — from prototype to a thing real people use.
Eric Green is a quantitative scientist and educator whose work focuses on how rigorous statistical methods can inform real-world decisions in health and medicine. At Duke he teaches and advises on causal inference, study design, and real-world evidence — including courses on AI in Global Health (GLHLTH 199) and Data Science & Data Visualization with R (GLHLTH 562).
He has also been a startup co-founder and chief science officer in digital health, and his work bridges academia, technology, and applied research — with an emphasis on transparency, rigor, and making complex analytical ideas usable in practice. He is the author of Global Health Research in Practice.
Research design, data science, or an AI product you want to get right. Let's talk about it.
[email protected]