Machine Learning Scientist- AI Imaging

About the Role

We are seeking a highly motivated Machine Learning Scientist with deep expertise in AI-driven image analysis to join our interdisciplinary team advancing precision therapeutics and diagnostics. You will lead the development and deployment of machine learning models that integrate histopathology imaging data (H&E-stained slides) with multi-modal biomarkers, driving clinical and scientific insights across large-scale cancer datasets.

You’ll work closely with scientists, pathologists, and engineers to innovate and apply deep learning approaches—including convolutional neural networks (CNNs) and vision transformers (ViTs)—to predict clinical outcomes, recurrence risk, and biological phenotypes from digital pathology images.

Responsibilities

• Design, implement, and evaluate deep learning models trained on high-resolution histopathology images.

• Develop multi-modal architectures that integrate H&E-stained slide data with genomic, transcriptomic, and ctDNA-based features.

• Contribute to the development of production-quality machine learning pipelines using AWS cloud infrastructure.

• Build scalable training workflows to handle tens of thousands of whole-slide images with automated pre-processing, tiling, and data augmentation.

• Collaborate with cross-functional teams including computational biologists, bioinformaticians, and clinical R&D teams.

• Communicate results through well-documented code, internal presentations, and peer-reviewed publications.

Required Qualifications

• PhD in Computer Science, Biomedical Engineering, Computational Biology, or a related field with a focus on machine learning or image analysis.

• Extensive experience in deep learning for medical imaging, especially digital pathology (H&E).

• Strong knowledge of CNNs, ViTs, attention mechanisms, and modern architectures for image understanding.

• Hands-on proficiency with PyTorch, Python, and SQL.

• Experience working with large datasets in cloud environments (preferably AWS).

• Demonstrated ability to independently drive ML research and translate it into production code.

• Experience using pre-trained foundation models for downstream finetuning.

Preferred Qualifications

• Experience with survival modeling, CoxPH loss, or time-to-event architectures.

• Exposure to self-supervised learning or weakly supervised MIL in whole-slide imaging.

• Familiarity with clinical oncology or biomarker discovery in cancer.

• Publications in top-tier conferences (e.g., MICCAI, CVPR, NeurIPS, ICML, ISBI) or journals (e.g., Nature Biomedical Engineering)