Salary
💰 $186,000 - $243,000 per year
About the role
- Develop digital pathology AI (DPAI) models with WSI data to predict clinical outcome and pathological/morphologic features, adapting open-source state-of-the-art AI foundation models to Veracyte’s data
- Evaluate and analyze DPAI models with respect to clinical, pathological and genomic outcomes, linking explainability of model features to biology
- Collaborate with internal and external partners, including medical and business development teams and key opinion leaders, to understand clinical and business requirements and tailor algorithms
- Work with bioinformaticians, statisticians, and medical experts to document projects and generate analyses and visualizations for peer-reviewed publication
- Provide technical leadership, independently drive project execution and completion, and communicate findings to experts and non-experts through presentations and academic writing
Requirements
- PhD in Data Science, Machine Learning, Applied Math or equivalent field
- 8+ years of experience in a data/applied scientist role or equivalent
- Expert in Python or equivalent language for AI/ML development, including data manipulation and preparation
- Proficient in statistical analysis, especially survival modelling and hypothesis testing (multivariate regression modelling with interaction effects)
- Experience working in cloud computing environments (AWS preferred)
- Demonstrated proficiency in summarizing and communicating findings from data with attention to detail
- Ability to work effectively in a fast-paced and collaborative environment
- Eagerness to learn new technologies and adapt to evolving requirements
- Ability to explain AI/ML concepts to experts and non-experts, including formal presentation and academic writing
- Experience with generation of publication-quality figures
- Knowledge of cancer biology (preferred)
- Proficiency with documentation and submission in regulated diagnostic environments (LDT or IVD) (preferred)
- Experience working with real world clinical data (preferred)