Open Postdoctoral position, faculty mentor Arden Morris

We seek a highly motivated computer scientist for a post-doctoral position in our multi-disciplinary health services research center within the Department of Surgery. The post-doc will have access to unique datasets, and can develop clinical partnerships across departments in Stanford's broader AI/ML ecosystem.  Stanford is a pioneer in in multi-disciplinary research, with its medical center sharing the same campus as arts and sciences.

The fellow will lead the development and validation of imaging-based models to predict patient response to cancer treatment (80%) and will manage the unit’s AI/ML core, ensuring robust, reproducible, and compliant pipelines across multidisciplinary collaborations (20%). The position offers protected time for independent research pursuing cross-cutting applications in surgical and cancer care, with formal mentorship and structured training to support future leadership in academic health data science.

Primary responsibilities

Design, implement, and validate predictive models that use imaging data (e.g., CT, MRI, PET) to forecast response to adjuvant treatments. Build end-to-end imaging pipelines: data ingestion, curation, preprocessing, segmentation/annotation support, feature extraction, model training, evaluation, and external validation. Collaborate with surgeons, radiologists, pathologists, biostatisticians, health services researchers, and health economists on clinical questions and study design. Contribute to grant writing, manuscript preparation, conference presentations, and dissemination of methods and findings. Support translational aims, including pilot deployments of predictive tools in clinical or health services settings and evaluation of real-world impact. Mentor research staff and junior trainees; foster cross-disciplinary collaboration and training within the core. Lead the AI/ML core’s operations, including governance of GPCs, reproducible workflows, version control, model deployment, monitoring, and documentation. Manage data privacy, security, and compliance considerations (e.g., HIPAA), data usage agreements, and governance for imaging datasets. Training plan and milestones.

Year 1: Establish data access streams, curate imaging cohorts, implement baseline predictive models, and create reproducible pipelines with documentation and version control. Develop a plan for cross-disciplinary collaborations and a personal research portfolio.

Year 2: Improve model robustness and generalizability with external validation, incorporate multi-modal data (e.g., imaging plus clinical variables), and pilot integration into decision-support workflows. Begin drafting grant proposals to expand the program.

Year 3: Lead an independent line of inquiry within the health services research unit, publish findings, and scale the AI/ML core to additional subspecialties. Demonstrate impact through at least one prospective or quasi-experimental evaluation of predictive tools in practice.

Environment and resources: We prioritize a collaborative environment spanning surgery, radiology, pathology, biostatistics, health services research, and health informatics. We provide access to high-performance computing, secure data environments, clinical imaging data, and software tools (Python, PyTorch/TensorFlow, radiomics, 3D Slicer, DICOM tooling). We foster a commitment to open science practices, rigorous validation, and pathways for translating methods into practice. 

ALTERNATIVE QUALIFICATIONS:

PhD in computer science, electrical/biomedical engineering, statistics, applied mathematics, or a related field. Strong track record in machine learning/deep learning with imaging data; publications in medical imaging or health data science are highly desirable. Proficiency in Python; experience with ML frameworks (PyTorch, TensorFlow); familiarity with imaging libraries, data governance, and version control (Git). Demonstrated ability to work across disciplines, communicate with clinicians, and manage projects with competing priorities. Experience with radiomics, multi-modal data integration, and ML Ops or deployment in clinical research settings.