I lead a team of researchers, data scientists, software engineers, technical managers, and software architects who are working to develop both scientific software and data analysis processes to enable IHME's researchers to perform cutting edge science in a more efficient, robust, and reproducible manner. We aim to identify statistical, computational, and data accessibility needs across the institute and then build tools and design processes to address them. Some of our projects include:- Scientific computing pipelines which enable large teams of researchers, analysts, and managers to collaborate on and automate complex data workflows- GUIs, APIs, data visualizations, and other tools which make our data, analytic techniques, and results more accessible to a wider audience, both internal and external- Advanced Bayesian modeling tools, which give even non-technical users the ability to fit large and sophisticated temporospatial models using techniques like algorithmic differentiation and Laplace approximation- A platform for easily building and executing large statistical simulations on top of Apache Spark- Microsimulation software that simulates the health and care of individuals within a population in far greater detail than ever before, enabling complex cost effectiveness analyses- Developing scientific computing best practices and helping teach and implement them across the institute- Early stage tests of migrating from on-premise to cloud-based computing clusters

I lead the Forecasting research team at IHME, which is extending the Global Burden of Disease study to include forecasts through 2040. These estimates cover over 300 causes of death and disability, nearly a hundred risk factors, and every country in the world (including many states and provinces) by age and sex.We are developing novel Bayesian statistical models to project what the world's health would look like if past trends and relationships continued into the future. In addition, we're building innovative simulation methods to examine how those forecasts might differ under alternative scenarios (e.g. what might the potential impact of a new shigella vaccine be? How might the number of lives saved differ under different rollout scenarios?). Importantly, we're doing so in a way that captures the complex interdependencies between different diseases, risk factors, and treatments.

I managed a team of data scientists, researchers, developers, and software architects who are trying to predict what the world's health is going to look like in 25 years and determine how governments and aid organizations can best improve it. We collected data, developed novel statistical models, built simulation software, engineered data pipelines, and created dynamic data visualizations to communicate our results.

At Symcat we have built an app that uses cutting edge diagnostic algorithms to give patients more accurate medical information. A recent study by Harvard researchers in the British Medical Journal ranked Symcat as the most accurate online diagnostic tool, beating out WebMD, the Mayo Clinic, and many others[1]. 1. https://www.washingtonpost.com/news/to-your-health/wp/2015/07/10/harvard-researchers-tested-23-online-symptom-checkers-most-got-failing-grades-heres-how-they-stack-up/

After my fellowship, I was promoted to IHME's first ever Data Scientist role. In this role I independently designed and built GBD Compare, about which Bill Gates said "in the whole world of data visualization this is one of the best efforts that has ever been done". [More info under projects]

IHME's Post Bachelor Fellowship program is a highly competitive research fellowship program at the intersection of global health and data science[1]. For my thesis, I developed an ensemble modeling platform for estimating deaths by country for every major disease in the world. This software has been used by nearly a hundred researchers for over 6 years now, providing the analytic basis for dozens of high impact peer-reviewed global health articles.1. http://www.healthdata.org/post-bachelor-fellowship

Every summer from 2002 through 2008 I worked in various capacities at Nemours Children's Health System. I started out volunteering in the clinical research department, doing basic analyses of survey data. By 2006 I worked directly for the Medical Director, building a reporting database of physician interactions. In 2008, under the guidance of the Chief Medical Informatics Officer, I developed a dynamic microsimulation of a new emergency room to predict staffing requirements.