At the Arnhold Institute, we seek to drive the next generation of Global Health impact through data-driven and open innovation. One of our core research goals is to discover the factors and their interactions that affect a person’s health outside the presence and absence of disease. We are seeking a probabilistic graphical models expert to model complex interactions between heterogeneous datasets (categorical, real-valued, unstructured text, etc.) Specifically, we are interested in modeling the interactions between large-scale “Big Data” (e.g. high-dimensional, regularly sampled, large sample sizes etc.) and small data (e.g. low dimensional, irregularly sampled, small sample sizes, etc.) Our ability to model and conduct inference on these complex interactions has the potential to be one of the greatest health-related discoveries.

Responsibilities

Work with domain experts to design physically-meaningful probabilistic graphical models that are both flexible and interpretable. The models must account for and reveal interactions between covariates.

  • Design various approximation and optimization techniques to get accurate and efficient predictions based on the designed models.
  • Design inference models to understand the impact changes on covariates have on target variables.
  • Design temporally evolving probabilistic graphical models.
  • Work with our human-computer interaction and data engineering teams to ensure the proper data are collected.
  • Design new model evaluation metrics that go beyond abstract general measures such as RMSE.
  • Ensure that all our data and results are reproducible and publicly available.
  • Lead and co-lead grant proposal preparation to advance the state-of-the-art of data-driven scientific discovery.

Qualifications

  • The successful candidate must be intellectually flexible with a proven track record of interdisciplinary work. We look for candidates who continuously step outside their comfort zone and work well with others outside their domain of expertise.
  • Algorithmic experience with large-scale predictive and inferential models. This is not a theoretical position. You must have computational experience to conduct complex probabilistic modeling at scale.
  • Experience with Bayesian modeling approaches such as Bayesian networks, dynamic Bayesian networks, and Bayesian program learning
  • Experience with experimental design and statistical analyses related to the generalizability of statistical models (e.g. reusable holdout).
  • Familiarity with spatio-temporal hierarchical models and other temporally-evolving graphical models
  • A Ph.D. in computer science, statistics, or applied mathematics.

To Apply…

Email the following documents to james.faghmous@mssm.edu with the title of the position in the email title (e.g. postdoctoral scholar, assistant professor, etc.)

  • Share the link to the source code for one large-scale graphical model project on Github
  • Share what you deem your most important publication/report with a paragraph why you chose that paper
  • A 2-page CV.
  • The names and contact information of three professional references.

About the Arnhold Institute for Global Health

Health equity is one of the greatest challenges of our times. We believe that access to quality healthcare for people and communities is a fundamental human right. The data science and product teams develop technologies to close the health inequity gap in the US and globally. Besides our mission, there are several reasons to consider joining us:

  • We are startup institute backed by a generous $25M gift from the Arnhold family
  • Your work will have the opportunity to be applied in NYC’s largest healthcare system and potential impact the well-being of millions of people.
  • We are a team of young and ambitious professionals from diverse cultures and backgrounds
  • The opportunity to build a transformative effort from the ground up and leave a legacy for those that will follow in your footsteps.
  • Access to world-class designers, system engineers, and health scientists to drive high-impact interdisciplinary work.
  • Travel to global settings to work with and get to know amazing collaborators and cultures.
  • We are located in the historic NY Academy of Medicine right next to Central Park on Fifth Avenue. Many of our team members conduct walking meetings in the Park!