Epidemiologist and biomedical data scientist. I analyze the dataset, build the narrative the evidence supports, and write the grant or paper that gets it funded and published. My maps have been used by a national Ministry of Health, my analyses published across 17 peer-reviewed papers, and my writing has won competitive research funding across my career.
Natural-history studies, registries, multi-omic and observational data. I find the signal and the structure that matter.
The figures, maps, and data story that turn an analysis into something a board, a reviewer, or a clinician acts on.
The grant, the paper, the presentation. Scientific writing that gets the work funded and published.
I’ve run this arc end to end: arbovirus risk maps used by the Colombian Ministry of Health, work published across 17 peer-reviewed papers, and competitive grants won from foundations and the NIH.
Most engagements start with one of these. They fit together, but each stands on its own.
I take a natural-history study, registry, or dataset and produce the analysis, the figures, and the narrative your audience will act on, ready for a board, a paper, or a presentation.
Grants, RFP responses, literature reviews, and manuscripts. I write the case for your science so it gets funded and published, drawing on a track record of winning the funding myself.
Natural-history protocols, registries, and analysis plans designed to answer the question and stand up to a reviewer, with geospatial and mapping methods where they add power. Built on leading field studies across three countries.
Gathering, harmonizing, and QC’ing messy multi-source datasets into something trustworthy and analysis-ready, the same work behind the paired genome and expression training data I built for VariantFormer at CZI Biohub. Deeper work, scoped and priced for the time it takes.
Domain-expert evaluation of a biological model’s outputs against ground-truth clinical and phenotype data. Grounded in building the training data for VariantFormer, a 1.2-billion-parameter model at CZI Biohub, and checking model predictions against patient phenotypes with clinicians.
Pre-investment scientific diligence for funds is available selectively, by referral, in domains where I hold direct expertise.
VariantFormer would not have been released if not for Amy’s efforts.
Her work strengthened the scientific credibility of VariantFormer and connected our models to real patients.
Amy is the best person to have on a tough project.
The kind of data scientist who makes the teams around her better.
Rare and ultra-rare disease is where I go deepest, as a foundation founder and the parent of a child with an ultra-rare syndrome. Small cohorts and thin evidence are the hardest version of every service above, and the one I know best.
My methods come from a decade in infectious-disease epidemiology and geospatial science: leading multi-country field studies, designing the data collection, and adapting and extending established spatial methods to new settings, the work behind most of those 17 papers. If your work lives in that world too, all the better.
I also build the structures research runs on: I’ve founded a 501(c)(3), convened a scientific advisory board across five institutions, and grown a 200+ family research network across 17 countries. When an engagement needs the collaboration built as well as the data analyzed, that’s in scope.
Tell me what you’re trying to land and the deadline you’re working against. It takes a minute, and it means our call starts with the real problem instead of warm-up.
Every engagement is independent and scoped in writing. I co-founded Artana Bio, a rare-disease company; where any engagement could touch its interests, I disclose it up front and recuse. The full conflict-of-interest policy is available on request.