Machine Learning Research Scientist

School of Medicine Established in 1930, Duke University School of Medicine is the youngest of the nation's top medical schools. Ranked sixth among medical schools in the nation, the School takes pride in being an inclusive community of outstanding learners, investigators, clinicians, and staff where interdisciplinary collaboration is embraced and great ideas accelerate translation of fundamental scientific discoveries to improve human health locally and around the globe. Composed of more than 2,500 faculty physicians and researchers, more than 1,300 students, and more than 6,000 staff, the Duke University School of Medicine along with the Duke University School of Nursing, Duke University Health System and the Private Diagnostic Clinic (PDC) comprise Duke Health. a world-class academic medical center. The Health System encompasses Duke University Hospital, Duke Regional Hospital, Duke Raleigh Hospital, Duke Primary Care, Duke Home and Hospice, Duke Health and Wellness, and multiple affiliations.

 

Position Summary 

This position is responsible for designing, developing, and validating machine learning products and systems within the scope of DIHI projects. This position will act as a liaison between product managers, data scientists, data engineers, and clinician users to determine the requirements for novel machine learning models. This position will assist with curation and quality assurance of data across diverse clinical and operational data sources and will collaborate closely with data engineers and solution architects to ensure that machine learning models are effectively integrated into clinical care. This position will work closely with the DIHI Program Director to identify and pursue opportunities to leverage emerging statistics and machine learning methods and technologies to drive impact within Duke Health. This role will participate in activities throughout the machine learning product lifecycle, including initial design, development, internal and external validation, temporal and clinical validation, impact and outcomes analyses, updating, and maintenance. This role will utilize statistics expertise to both develop machine learning systems as well as assist in the design and conduct of evaluation studies to measure the effect of algorithms integrated into clinical care. Lastly, this position will help teach students and trainees involved in DIHI workforce development programs. 

 

This position will possess expertise in statistics and machine learning and causal inference to effectively develop machine learning solutions to improve care at Duke Health and beyond. This role will work closely with business unit leaders and compliance and regulatory support staff, and success carrying out responsibilities will be strengthened by strong interest in business, ethics, law, and policy. This position will maintain comprehensive and contemporary knowledge of machine learning and statistics methods and best practices, with a focus on development and validation of products and systems in the health setting. This position reports directly to the DIHI Program Director. 

 

Duties and Responsibilities of Position: 

 

• Development, validation, and implementation of machine learning systems

• Works directly with data science lead, quantitative science trainees, and clinical trainees who are involved in model development and validation 
• Develops utilities and frameworks to standardize and enhance the development and validation of machine learning systems 
• Develops and updates monitoring systems to ensure that any implemented machine learning model continues to perform as specified 
• Reviews code and performs evaluations of previously built machine learning models before integration into technical infrastructure 
• Maintains contemporary knowledge of advances in machine learning to ensure that emerging technologies are incorporated into DIHI development efforts 

 

• Evaluation, monitoring, maintenance, and oversight of implemented machine learning systems 

• Works with clinical and operational leaders to define relevant outcome measures to demonstrate impact of machine learning products integrated into care 
• Designs and conducts evaluations of machine learning products to identify strengths and weaknesses of model integrations and improve practices across DIHI in regards to machine learning model development 
• Collaborates directly with data engineering and solution architect team members to optimize technology infrastructure for rapid implementation of machine learning systems 
• Defines model performance measures related to safety, efficacy, equity, and fairness to monitor and measure throughout the product lifecycle as well as standard reporting practices to diverse stakeholders 

• Identifies opportunities