Join Optum, a global leader in health optimization, and contribute to improving health outcomes worldwide. We are seeking a skilled ML Engineer to design and build robust ML/LLM inference platforms on AWS and Azure. This role is integral to connecting individuals with the care and resources they need, fostering healthier lives.

At Optum, you'll be part of an inclusive culture, collaborating with talented peers and accessing comprehensive benefits and development opportunities. Make a tangible impact as we advance health optimization on a global scale, embodying our commitment to Caring, Connecting, and Growing together.

Key responsibilities include developing and maintaining production ML/LLM inference pipelines, deploying and managing ML workloads on Kubernetes (EKS, AKS), and building scalable inference services using vLLM and REST APIs.

You will also design event-driven architectures with queues and messaging services, and establish CI/CD pipelines for ML models and inference. This involves managing model versioning, deployments, and releases, alongside optimizing inference for GPUs, latency, throughput, reliability, and cost. Collaboration with Data Scientists, Platform/Cloud teams, and leadership is essential for architectural decisions and delivery.

Adherence to all company policies, including those related to work location, team assignments, schedules, and flexible work arrangements, is required.

We are looking for a seasoned ML Engineer with expertise in Python, ML/LLM Inference, AWS, Azure, and Kubernetes (EKS/AKS).

Essential skills encompass Docker, vLLM, REST APIs, CI/CD practices, Event-Driven Architecture, GPU workloads, and MLOps. A strong understanding of model versioning, deployment strategies, and inference optimization is crucial.

This role requires the ability to review peer code and designs, champion engineering best practices, and work closely with cross-functional teams. A commitment to advancing equitable care and addressing health disparities, aligning with UnitedHealth Group's mission, is highly valued.