Data Scientist - Medical Imaging AI

As a Data Scientist, you will be a key contributor to our cutting-edge research initiatives in Cancer Imaging. You will play a leading role in developing multimodal foundation models for downstream tasks such as classification, segmentation, and synthetic image/text generation. Starting with mammography, the successful candidate will work with diverse imaging modalities, including Chest X-rays, CT, and MRI, to address specific challenges in each domain.

Responsibilities:

Model Development:
-Design generalist pre-trained models for medical imaging using strategies like transfer learning, contrastive training, masking and self-supervised modeling.
- Implement and experiment with architectures such as CNNs, EfficientNet, and Transformers.

Multimodal Understanding:
- Develop large language models and large vision models to investigate representation learning, visual question answering and visual grounding.
- Focus on learning joint representations from imaging data and clinical reports for comprehensive understanding.

Data Preprocessing:
- Prepare and preprocess medical images using Python and relevant libraries (e.g., pandas, NumPy, OpenCV).
- Handle data augmentation, normalization, and other preprocessing steps to enhance model performance.
- Maintain clear and comprehensive documentation for models, datasets, and preprocessing pipelines.

Collaboration:
- Collaborate with domain experts to ensure accurate and meaningful labeled datasets for training and evaluation.
- Participate in discussions to understand clinical requirements and adapt models accordingly.
- Collaborate with radiologists, data scientists, and software engineers to align research objectives with real-world medical imaging needs.