Senior Applied AI Scientist
At PinkStar, we’re on a mission to help detect breast cancer earlier, when it can make the biggest difference. We’re developing advanced AI for MRI to uncover imaging biomarkers and recognize subtle tissue changes before lesions form, turning those insights into clinical workflows that can scale. This is an opportunity to take meaningful scientific ownership in an early-stage company, collaborate closely with clinical experts, and build technology that could redefine early detection, making diagnosis earlier, smarter, and more personalized.
We are seeking a hands-on Applied AI Scientist to lead the development of our core deep learning models, from initial concept through achieving industry-grade robustness and clinical reliability. This is a model-centric role focused on refining and maturing algorithms that must perform consistently across real-world clinical data. You will work closely with our scientific and clinical team to ensure models capture meaningful biological signals and translate into clinically valuable tools, while helping shape the company’s technological foundations. As an early-stage startup, this role requires a flexible and entrepreneurial mindset, and willingness to engage with broad technical challenges to advance our core technology.
Key Responsibilities:
- Lead the design, development, and validation of advanced deep learning models for breast MRI, including computer vision tasks such as detection, segmentation and classification, with a focus on robust and clinically reliable performance for commercial products.
- Drive model innovation through extensive experimentation with modern computer vision and deep learning approaches, including 2D/3D CNNs, transformers and generative techniques, and advanced training strategies.
- Design, implement, and optimize MRI preprocessing pipelines, including registration, normalization, harmonization, quality control, to support reliable and reproducible model development.
- Develop models that are robust and generalizable across large, heterogeneous datasets, imaging protocols, scanners, and clinical environments.
- Analyze model behavior, identify performance limitations and failure modes, and systematically improve model accuracy, robustness, and reliability.
- Work closely with radiologists and clinical collaborators to ensure models capture meaningful biological and clinical signals and translate into clinically useful tools.
Required qualifications:
- MSc/PhD in Computer Science, Biomedical Engineering, Electrical Engineering or a closely related field.
- 3+ years of strong hands-on experience developing deep learning models for medical imaging in a commercial company, with deep expertise in computer vision and using modern architectures such as convolutional neural networks.
- Solid understanding of medical image processing, medical data workflows, and deep learning model evaluation methodologies.
- Ability to work independently, take ownership of model development, lead multiple projects, and operate effectively in a fast-moving startup environment.
- Strong experience with deep learning frameworks such as PyTorch or TensorFlow.
- Excellent communication skills and comfort working in a multidisciplinary team.
Preferred qualifications:
- Experience with multi-site datasets, harmonization and model generalization challenges.
- Familiarity with federated learning, privacy-preserving training methods, and clinical trial workflows.
