Postdoctoral Researcher Position-UCSF Radiation Oncology
Postdoctoral Researcher – Personalized Brain Metastasis Risk Modeling
Institution: UCSF Department of Radiation Oncology (Parnassus Campus)
Location: San Francisco, CA
Start Date: Flexible, position available immediately
Funding: Full time, NIH R01 grant support (5 years)
Overview
We are seeking a highly motivated Postdoctoral Researcher to join our interdisciplinary team developing personalized, time-based voxel-wise risk models of brain metastases (BM) formation. This NIH-funded project integrates medical imaging, deep learning, and quantitative modeling to improve patient-specific prediction of BM occurrence and support novel radiation treatment planning approaches.
The postdoc will have the opportunity to work at the intersection of medical physics, computational imaging, oncology, and AI, contributing to a project with direct translational relevance for improving outcomes in patients with non-small cell lung cancer (NSCLC) and brain metastases.
Key Responsibilities
-
Model Development:
- Design and implement Bayesian deep learning architectures (e.g., Bayesian UNET 3D within MONAI/PyTorch) to integrate multimodal imaging data (anatomy, perfusion, functional tracks) with population BM risk maps.
- Develop time-based prediction models incorporating lesion temporal history for longitudinal risk estimation at 3, 6, and 12 months.
-
Data Integration & Preprocessing:
- Work with a multi-institutional NSCLC BM imaging database (https://www.nature.com/articles/s41467-025-59584-7) including MRI (anatomical and perfusion), BM contours, and clinical annotations.
- Perform advanced image registration (rigid + deformable) to map lesions and perfusion data onto MNI brain space and transform back into patient space.
- Automate anatomical and functional brain segmentation (e.g., FastSurfer, HCP atlas integration).
-
Model Evaluation:
- Quantitatively assess prediction accuracy, uncertainty, and robustness using cross-validation and longitudinal validation cohorts.
- Employ advanced hyperparameter optimization strategies (e.g., Optuna) and Bayesian inference methods to enhance reliability.
-
Collaboration & Dissemination:
- Work closely with physicists, oncologists, radiologists, and biostatisticians across multiple institutions.
- Publish high-impact scientific articles and present at major conferences (AAPM, ASTRO, MICCAI, IBSI).
- Contribute to mentoring graduate students and residents in imaging and AI research.
- Participation to 1-2 conference per year for development, networking and learning opportunity
Qualifications
Required:
- PhD in Medical Physics, Biomedical Engineering, Computer Science, Applied Mathematics, or related field.
- Strong background in medical image analysis, machine learning/deep learning, or computational modeling.
- Proficiency in Python and deep learning frameworks (PyTorch, MONAI, TensorFlow).
- Experience with neuroimaging (MRI registration, segmentation, perfusion analysis) and 3D data handling.
Preferred:
- Prior research in oncology, radiation therapy, or neuroimaging applications.
- Experience with Bayesian modeling, uncertainty quantification, and survival analysis.
- Familiarity with multi-institutional datasets, data harmonization, and reproducible workflows.
- Strong publication record and evidence of independent research capability.
Opportunities & Career Development
- Work on a seminal NIH-funded project with a unique multi-institutional dataset.
- Access to cutting-edge computational infrastructure and clinical imaging resources.
- Mentorship and career development support from an interdisciplinary team of investigators.
- Potential to transition into independent funding opportunities (e.g., NIH K99/R00, early faculty positions).
Application Instructions
Applicants should submit:
- A cover letter detailing research experience, interests, and career goals.
- Curriculum vitae (CV).
- Contact information for 2-3 references.
Send applications or inquiries to:
Olivier Morin, PhD
Principal Investigator
Olivier.Morin@ucsf.edu