Machine Learning Scientist, Multimodal AI

POSITION SUMMARY:

Natera is hiring a Machine Learning Scientist to join our AI and computational biology team. This role develops and deploys deep learning models across digital pathology, genomics, transcriptomics, and cell-free DNA (cfDNA) modalities. You will build multimodal AI systems that integrate imaging, molecular, and clinical data, leveraging proprietary genomic and clinical datasets. You will collaborate with scientists, pathologists, bioinformaticians, and software engineers to scale machine learning approaches that advance personalized oncology diagnostics and tumor-informed minimal residual disease (MRD) testing.

PRIMARY RESPONSIBILITIES:

Design, implement, and evaluate deep learning models across biomedical data modalities, including histopathology imaging, genomic sequencing, transcriptomics, and cfDNA features

Develop multimodal AI architectures that integrate H&E whole-slide imaging data with molecular and clinical data sources

Build scalable, production-quality machine learning workflows and pipelines using cloud infrastructure (AWS)

Apply modern machine learning techniques including convolutional neural networks (CNNs), vision transformers (ViTs), sequence transformers, representation learning, and foundation model fine-tuning

Collaborate across technical and clinical teams to translate machine learning prototypes into validated tools

Analyze model outputs to generate reproducible biological and clinical insights

Document pipelines thoroughly and communicate data-driven findings clearly to cross-functional stakeholders

QUALIFICATIONS:

PhD in Computer Science, Computational Biology, Biomedical Engineering, Bioinformatics, Statistics, or a related quantitative discipline with a focus on machine learning or AI

Core experience developing machine learning models for biomedical applications, specifically in medical imaging, computational pathology, genomics, transcriptomics, multi-omics, or molecular diagnostics

Hands-on expertise with PyTorch and strong production-level programming skills in Python

Practical application of deep learning architectures such as CNNs, transformers, attention mechanisms, and representation learning

Experience managing datasets and training workflows within distributed or cloud computing environments (AWS)

Proven ability to take ownership of research projects and translate prototypes into robust, deployment-ready workflows

Experience adapting pre-trained foundation models for downstream biomedical applications

PREFERRED QUALIFICATIONS:

Experience integrating imaging, molecular, and clinical data within unified multimodal machine learning frameworks

Technical familiarity with DNA sequencing, RNA sequencing, methylation, and ctDNA assays

Hands-on experience with digital pathology software and whole-slide imaging analysis

Exposure to survival modeling, longitudinal prediction, or time-to-event modeling

Experience applying self-supervised learning, weakly supervised learning, or multiple instance learning (MIL) to clinical data