ARUP Laboratories is looking to improve the accuracy and efficiency of clinical labs through data-driven methods, aiming to advance the state-of-the-art in clinical diagnosis, treatment, data analysis, and reporting.
Requirements
- Advanced understanding of multiple machine learning approaches (supervised and unsupervised, classification techniques, text and image analysis)
- Proficiency working in linux / unix data analysis environments
- Comfort working in large, complex codebases
- Proficiency with modern cloud computing and HPC environments for machine learning
- Ability to summarize complex, large data sets into clear, interpretable results
- Six years’ experience developing machine learning tools with one or more machine learning frameworks (Scikit-learn, PyTorch, JAX, Tensorflow, etc).
- Six years’ experience working with one of the following programming languages: Python, Java, C-Sharp, Javascript, R or similar
Responsibilities
- Creates, refines and validates software, data analysis processes, statistical, and data visualization tools.
- Implements and maintains computational tools for data mining and data integration.
- Develops enhanced data collection procedures to include information relevant for building analytic systems
- Develops new tools to aid in the analysis of assay quality metrics, interpretation, and reporting.
- Develops novel methods/algorithms, and extends, adapts and improves existing methods as needed.
- Develops analysis approaches in support of clinical applications.
- Works effectively with other scientists, programmers, engineers, and researchers to implement novel or adapt existing approaches to data analysis and reporting.
Other
- Ability to communicate effectively in English
- Ability to collaborate in a cross-functional environment
- Communicate results to non-technical staff and faculty in a clear, compelling manner.
- Masters’ Degree in a quantitative science or relevant technical field or equivalent experience
- Ability to write clear, well-reasoned text descriptions of methods, suitable for publication
