Postdoctoral Associate

• Organize, curate, and maintain large-scale ophthalmic imaging datasets, including OCT, OCTA, and ultra-widefield fundus imaging.
• Interface with imaging platforms and data export systems to ensure efficient data acquisition.
• Maintain accurate file naming conventions, metadata standards, dataset version control, and audit trails.
• Build and manage structured databases linking imaging data with clinical and research datasets.
• Perform data cleaning, standardization, and quality assurance activities to support reproducible research.

• Preprocess ophthalmic images for machine learning applications through resizing, normalization, and artifact reduction.
• Coordinate image annotation and segmentation efforts with clinicians and research staff.
• Develop and maintain annotation standards, documentation, and quality control processes.
• Ensure consistency and accuracy across imaging datasets and labeling workflows.

• Support the design, development, and optimization of machine learning and deep learning models.
• Prepare datasets for training, validation, and testing while maintaining rigorous data integrity standards.
• Execute model training workflows and monitor model performance.
• Conduct hyperparameter optimization and experiment tracking.
• Collaborate with research teams to implement robust validation strategies and statistical methodologies.

• Evaluate model performance using industry-standard metrics, including AUC, sensitivity, specificity, precision-recall analysis, and ROC curves.
• Assess model generalizability across diverse research cohorts.
• Identify potential sources of bias, overfitting, and data leakage.
• Develop performance visualizations and communicate findings to stakeholders.

• Integrate imaging-derived biomarkers with clinical, genetic, and cognitive datasets.
• Support statistical analyses and exploratory research initiatives.
• Generate tables, figures, and analytical summaries for publications and presentations.
• Translate complex analytical outputs into clinically relevant insights.

• Build scalable, reproducible workflows for data preprocessing, training, validation, and evaluation.
• Automate routine data management and quality assurance processes.
• Utilize version control and best practices for code management and documentation.
• Optimize computational efficiency across research computing environments.

• Support compliance with institutional policies, research protocols, and applicable privacy regulations.
• Assist with data de-identification and secure data management practices.
• Maintain documentation supporting audits, protocol requirements, and research reproducibility.

• Work closely with multidisciplinary research teams across clinical and scientific domains.
• Present project updates and findings during laboratory and research meetings.
• Document workflows and methodologies to facilitate reproducibility and knowledge sharing.
• Mentor and support trainees, students, and new team members as appropriate.

• Assist with manuscript development, abstracts, presentations, and grant-related research activities.
• Create scientific figures, model visualizations, and technical documentation.
• Contribute to methods development and reporting for peer-reviewed publications and conference submissions.

• Experience implementing novel machine learning architectures and emerging AI methodologies.
• Knowledge of multimodal modeling approaches integrating imaging, clinical, and genetic data.
• Familiarity with model explainability techniques such as saliency mapping and Grad-CAM.
• Ability to contribute to study design, hypothesis development, and innovative research initiatives.

Duke is an Equal Opportunity Employer committed to providing employment opportunity without regard to an individual's age, color, disability, gender, gender expression, gender identity, genetic information, national origin, race, religion, sex (including pregnancy and pregnancy related conditions), sexual orientation or military status.