Postdoctoral Scholar Position in Trauma, Injury and Inflammation Prediction Modeling

University of Washington Tacoma
School of Engineering and Technology
United StatesWashingtonTacoma
apply.interfolio.com/111895

Description

The University of Washington has an outstanding opportunity for a full-time postdoctoral researcher in AI driven Trauma, Injury and Inflammation Prediction Modeling.

We are looking for an enthusiastic, self-motivated individual who has experience with machine learning algorithms and who is passionate about advancing the science behind predicting onset of sepsis for critical care and trauma patients.

Candidates should have experience with research using cloud-based ML modeling tools, especially TensorFlow and PyTorch, and handling healthcare datasets. Candidates should see value both in excellent scholarship and in real-word impact.

Our group uses predictive models to (1) understand the temporal variables that lead to onset of sepsis and sepsis shock, and (2) to test specific hypothesis that could help address those underlying factors. For a current project, which is a collaboration between Department of Surgery and the Center for Data Science we are looking for a postdoctoral researcher to join our team of AI/ML experts, Surgeons, and doctoral students. We work with Harborview Medical Center team to understand questions and potential solutions for classifying and recommending actions that can help reduce mortality amongst the trauma patient populations. We also aim to make the models more useful and usable in the clinical setting; to do so, we will implement, and improve the usability of, the machine learning models (including deep learning models) we have currently developed.

In this project, testing potential solutions means understanding how new cohorts with different inclusion exclusion criteria can be setup to define the target variables, how changes in different types and granularity of features can shape the predictive modeling process and how techniques such as deep learning and graph-based models can help improve the state of the art of this very severe and complex problem.

This position is initially offered for 1 year (12-month) 100%-time FTE, with renewability expected, depending on job performance and funding. The preferred start is Fall 2022. The postdoc will be hosted in the UW Center for Data Science (dscience.tacoma.uw.edu), at UW Tacoma, supervised by Professors Grant O’Keefe, Ankur Teredesai, and Juhua Hu. Collaborators include faculty colleagues at UW Medicine and UW Harborview Medical Center.

Salary and other support for this position are provided through an institutional Post-doctoral training grant from the National Institute of General Medical Sciences (Institutional Postdoctoral National Research Service Award (NRSA - T32) awarded to the UW Department of Surgery. Trainees receive at least 2 years of research training in basic, clinical and/or translational science areas and engage in activities to promote research career development. Additional information and the online application can be found here: uwsurgery.org/divisions/trauma-and-burn/research/.

Duties and Responsibilities:
The successful candidate will conduct excellent scholarly research and will support ongoing efforts in sepsis prediction. Primary duties & responsibilities include design and development of various prediction models, using these models to evaluate different scenarios in terms of their impact on trauma related sepsis or septic shock, and engaging with care management team to test the efficacy of developed models. Additional duties & responsibilities include writing and editing peer-reviewed journal articles, communicating with diverse collaborators and stakeholders, presenting at scientific conferences and to community groups, and brainstorming ways to use scientific analyses to pose and test solutions to sepsis prediction.

Postdoctoral scholars at UW are represented by UAW 4121 and are subject to a collective bargaining agreement, unless agreed exclusion criteria apply. For more information, please visit the University of Washington Labor Relations website: hr.uw.edu/files/labor/UAW-4121-Postdoc-2021-2023-CBA-TA.pdf


Qualifications

Minimum Qualifications:
Specific knowledge, skills and abilities required include:
• A PhD (at the time of appointment) in science, math, or engineering;
• Able to develop, maintain, and apply deep learning frameworks to solve complex data problems in cloud computing environments;
• Able to process datasets using Python and SQL;
• Experience and/or demonstrated interest in deep learning for healthcare research, time series, and graph neural networks;
• Excellent verbal, written, and interpersonal communication skills, with record of high-quality publications in peer-reviewed scientific journals;
• Passionate, self-motivated and able to work independently and collaboratively.

The first qualification above (“ability to develop, maintain, and apply deep learning frameworks to solve complex data problems in cloud computing environments”) includes abilities such as (1) data analysis and data-file management steps; for example, ability to handle raw data and transform it into features, overlaying logical common data models that can be reused by others, computing and maintaining experiments and model results and familiarity with engineering workflows for code development; (2) investigating the mathematical representations of data coming from clinical domains, (3) learning how to convert outputs from ML frameworks into scoring pipelines, and (4) working collaboratively with others to set up code-reviews and investigate model performance and outputs.

Desired Qualifications:
Abilities that are desirable, providing for an enhanced level of job performance, include these:
• Comfort and familiarity with running scientific models in a cloud computing environment;
• Experience with community engaged scientific research, with a preference for healthcare and clinical datasets;
• Familiarity with large domain specific datasets (not necessarily healthcare) and/or with datasets or models specific to extracts from commercial electronic health records and/or claims data;
• Programming expertise in python.
(Those skills are desirable but not required. We welcome candidates that have one or more of those skills; we do not expect that candidates would have all of the skills on this “desirable” list.).


Start date

To be determined

How to Apply

To apply, applicants should send the following application materials to ankurt@uw.edu with the subject “your name: Postdoc in ML for Sepsis Application”.

Applications will be reviewed starting August 30, 2022.
A cover letter (2 pages max) briefly describing qualifications, research interests, professional goals, and specific interest in this position. If you are missing any of the "minimum qualifications", or if you have any of the "desired qualifications,” please mention this in your cover letter.
A research goals statement (1 page max) describing the candidate’s personal and professional experiences and their contributions to and vision for advances in machine learning and healthcare.
Resume or CV.
Contact information (name, position, email) for 2 to 4 references.
Sexual Misconduct form: State law requires that the University of Washington obtain the Disclosure of Sexual Misconduct declaration signed by the candidate. The declaration will require you to disclose any substantiated findings of sexual misconduct, to authorize current and past employers to disclose to the UW any sexual misconduct currently being investigated and/or committed by you, and to release current and past employers from any liability.

Please include your name as part of the filename.
Position URL: apply.interfolio.com/111895


Contact

Ankur Teredesai
ankurt@uw.edu