Senior Principal AI​/ML Engineer - Research Sovereign AI

Why Mayo Clinic

Mayo Clinic is top-ranked in more specialties than any other care provider according to U.S. News & World Report. As we work together to put the needs of the patient first, we are also dedicated to our employees, investing in competitive compensation and comprehensive benefit plans – to take care of you and your family, now and in the future. And with continuing education and advancement opportunities at every turn, you can build a long, successful career with Mayo Clinic.

• Medical:
  Multiple plan options.
• Dental:
  Delta Dental or reimbursement account for flexible coverage.
• Vision:
  Affordable plan with national network.
• Pre-Tax Savings: HSA and FSAs for eligible expenses.
• Retirement:
  Competitive retirement package to secure your future.

The Senior Principal AI/ML Engineer for AI Representation & EMR Vectorization is the senior technical leader and lead scientist responsible for architecting Mayo Clinic’s unified multimodal EMR representation layer. This role defines and builds the scientific substrate used by foundational models, clinical agents, and research applications. The individual serves as a hands‑on expert and player‑coach, guiding technical strategy while contributing directly to model development, graph construction, and representation science.

Over time, this position will build and lead a specialized team.

• Design and implement Mayo’s multimodal EMR representation AI architecture, including text, imaging, waveform, structured data, temporal sequences, and multi‑visit trajectories.
• Develop graph-based representations and knowledge graphs linking patients, events, attributes, clinical concepts, and embeddings.
• Integrate graph reasoning, vector similarity search, and hybrid vector–graph pipelines for retrieval‑augmented models and agentic reasoning.
• Define standards for temporal modeling, drift‑aware embeddings, and sequence alignment across encounters.

• Build large‑scale embedding pipelines using transformer‑based models, contrastive learning, graph neural networks, and hybrid architectures.
• Implement efficient query layers using vector stores and graph databases.
• Develop interpretable embedding diagnostics, attribution tools, and graph‑based audits to enable safe clinical use.
• Explore and implement methods for explaining similarity, graph traversals, temporal evolution, and patient‑neighbourhood reasoning.

• Work with AI researchers on specialty‑specific embeddings, representation refinement, and research prototypes.
• Collaborate with clinicians to operationalize clinically meaningful features, phenotypes, and longitudinal concepts.
• Provide scientific input to the Foundational Model Science Program to ensure alignment between representations and model architectures.

• Serve as founding technical lead of the Reasoning EMR Representation team.
• Mentor junior scientists and engineers; build a future team specializing in representation learning and graph‑based reasoning.

• Master’s in Computer Science, Machine Learning, Biomedical Engineering, or related field. 9 years of relevant experience, or a bachelor’s degree with 11 years of relevant experience.
• Extensive (9+ years) experience applying AI and machine learning in production healthcare environments or similar highly regulated or technology focused industries, showcasing an acute understanding of healthcare technology.
• Hands‑on expertise with graph databases, and knowledge graph construction.
• Strong experience with transformer‑based models, contrastive learning, and temporal modeling.
• Experience designing or deploying vector search systems and hybrid vector–graph reasoning pipelines.

• PhD or Master’s in Computer Science, Machine Learning, Biomedical Engineering, or related field.
• 10+ years experience building production ML systems, including multimodal architectures and representation learning.
• Experience with EMR data, healthcare multimodality, or clinical data integration.
• Experience building patient similarity models, temporal embedding…