Data Scientist, AI Biomedical Imaging

Job Summary

The mission of Novartis is to reimagine medicine, and our team advances that mission by applying advanced image analysis, computer vision, and AI methods to early drug discovery. We partner closely with experimental scientists, disease‑area teams, data scientists, bioinformatics experts, and platform engineers to extract meaningful biological insight across diverse imaging modalities (high‑content screening, custom microscopy platforms) and biological model systems (cellular assays, co‑cultures, organoids, tissue models).

To grow this capability, we are seeking a seasoned, innovative, and collaborative data scientist with deep expertise in AI‑enabled image analysis to join the Data Science team in Discovery Sciences (DSc) at Novartis Biomedical Research, Cambridge, MA.

Internal

Job Title:

Senior Expert I/II, Data Science

Location:

Onsite, Cambridge, MA

• Lead AI‑enabled image analysis strategies for complex biological imaging workflows, acting as the embedded imaging‑AI scientific partner working side‑by‑side with wet‑lab scientists to understand emerging assay needs, align approaches with scientific priorities and platform standards, and explain advanced AI concepts in accessible terms.
• Identify high‑impact opportunities where AI can deliver meaningful scientific value and define rigorous benchmarking and evaluation strategies to guide method selection.
• Develop, validate, and deploy robust image analysis algorithms to characterize cellular, organoid, tissue, and other complex biological phenotypes in high‑throughput and high‑content imaging data, generating reproducible outputs that support decision‑making in drug discovery projects.
• Drive adoption of advanced AI methods for imaging, including deep learning, vision foundation models, embedding‑based phenotyping, segmentation, classification, and multimodal integration, translating state‑of‑the‑art methods into practical, validated workflows that augment expert review and enable scalable interpretation of large, high‑dimensional datasets.
• Contribute to scalable, reusable image analysis workflows in partnership with other data scientists, data engineering, and platform teams, championing best practices across the workflow lifecycle.

• PhD in computer science, AI, machine learning, biomedical image analysis, computational imaging, data science, or a related quantitative field, with 3+ years of applied experience in AI for bioimaging and computer vision.
• Demonstrated experience developing and validating image analysis algorithms for biological, biomedical, or pharmaceutical research applications, with practical experience in image segmentation, feature extraction, phenotypic profiling, object classification, or representation learning applied to high‑content or high‑throughput imaging data.
• Practical expertise in designing benchmarking and evaluation strategies to compare image analysis methods and guide rigorous, evidence‑based model selection.
• Ability to work effectively in Linux‑based high‑performance computing, cloud, or large‑scale data processing environments, with a strong commitment to reproducible research, version control, testing, and data provenance.
• Strong proficiency in Python and the scientific deep learning stack (e.g., PyTorch, Hugging Face, Lightning, MONAI), along with hands‑on experience using image analysis tools such as scikit‑image, OpenCV, napari, Cellpose, Star Dist, Instan Seg, and OME‑Zarr.
• Self‑motivated experienced contributor who thrives in a collaborative, multidisciplinary environment with biologists, imaging scientists, software engineers, and bioinformatics partners, working with appropriate independence and helping shape project direction through both technical expertise and scientific judgment.
• Excellent scientific communication and stakeholder engagement skills, including the ability to explain complex AI and image analysis concepts to experimental scientists, project teams, engineers, and non‑technical audiences.

• Experience developing or adapting foundation models, self‑supervised learning approaches, multimodal AI models, or embedding‑based analysis…