Software Engineer (MLOps) – Early Disease Detection Clinical AI

Lucem Health, created by Mayo Clinic in 2021, aims to revolutionize care delivery through clinical AI. As a Software Engineer (MLOps), you will bridge the gap between clinical machine learning and production-grade software engineering, contributing to automated infrastructure and deployment pipelines to enhance patient outcomes.

Responsibilities:

• Collaborate with clinical experts, data scientists, and software developers to translate business and clinical needs into robust, scalable MLOps solutions
• Design, build, and optimize end-to-end automated pipelines for clinical model training, retrospective validation, deployment, and monitoring
• Work with data scientists to implement automated model evaluation, benchmarking, and retraining logic, ensuring deep alignment between model performance metrics and production guardrails
• Collaborate with the Data Engineering team to define and consume standardized datasets from common clinical data models (e.g., OMOP), ensuring models are fed with highly structured, clean clinical data
• Maintain and expand our automated model deployment pipeline, including all model, code, data artifacts, workflows, and repositories, to enable seamless, “push-button” production deployments and automated retrospective validations
• Implement automated monitoring, logging, and alerting systems to track model inputs, output feature drift, and operational latency in Google Cloud Platform (GCP) production environments
• Manage and optimize cloud infrastructure for machine learning workloads specifically on Google Cloud Platform (GCP)
• Develop and maintain infrastructure as code using tools like Terraform
• Create and maintain clear architecture diagrams to document, justify, and communicate decisions for new MLOps infrastructure and patterns
• Write and maintain highly performant, production-grade applications and automation scripts for core MLOps services
• Contribute to robust clinical model governance practices, including drafting and automating model documentation templates (such as CHAI Model Cards) to track clinical bias, model performance parameters, and data drift over time
• Stay updated with the latest trends in MLOps, deployment patterns, and healthcare-focused model governance