GSK aims to develop transformational medicines using genetics, functional genomics, and machine learning, with AI playing a role in diagnosing and using medicines to improve health outcomes. The AI/ML Medical Imaging Team specifically focuses on applying machine learning and AI, particularly deep learning, to medical imaging data (CT/MRI/PET) to extract insights for novel biomarker/phenotype and imaging endpoint development in clinical trials.
Requirements
- Familiarity with current deep learning literature and math of machine learning
- Experience in software engineering with advanced skills and expertise in best practices for Pythonic programming, for example refactoring code for efficiency and modularization in PyTorch.
- Developing, troubleshooting, and validating deep learning models for diverse computer vision tasks (e.g., segmentation, recognition, classification, domain adaptation) using PyTorch.
- Experience in one or more open-source computer vision libraries such as Pillow, scikit-image, and OpenCV.
- Experienced in working with clinical imaging data, especially osteoarthritis or respiratory imaging
- Knowledge in disease biology.
- Track record of projects or peer-reviewed publications at the intersection of computer vision and medical imaging.
Responsibilities
- Carry out product-driven research on novel machine learning methods to medical imaging data.
- Leverage internal high performance computing system to train and productionize our models at scale.
- Work closely with domain experts on cross-disciplinary teams to generate actionable insights that impact phenotype, biomarker and clinical endpoint development.
- Contribute to our developing codebase with well-tested, production-ready code.
Other
- PhD in computer science, engineering, math or biological sciences.
- Track record of writing software in a team in industrial environments or open-source projects.
- Mentality of commit early and often, metrics before models, and shipping.
- Competitive candidates will have in-depth knowledge of machine learning with a track record of developing machine learning and especially deep learning models for solving challenging real world scientific problems.
- They should be comfortable with writing quality, well-documented, and well-tested code in the AI/ML space and operate in an agile environment.
