Lead AI/ML Engineer - Office-Based in Eden Prairie, MN

UnitedHealth Group•Eden Prairie, MN

3d•$145,500 - $249,500•Onsite

About The Position

Optum Tech is building agentic AI systems to transform healthcare operations through autonomous and human-in-the-loop agents that reason over healthcare use cases. The Lead AI/ML Engineer will design, build, and deploy production LLM-powered agents that make decisions on claims, helping reduce cycle time and manual work. This is a hands-on engineering role with end-to-end ownership of agentic workflows, from problem framing and orchestration design through production monitoring. The position follows an office-based schedule with four in-office days per week.

Requirements

Undergraduate degree or equivalent experience.
At least 5 years of experience building and deploying ML systems in production.
At least 5 years of experience with Python.
At least 5 years of software engineering experience.
At least 3 years of experience with LLMs, including prompting, RAG, and fine-tuning.
At least 1 year of experience with agentic AI systems.

Nice To Haves

Experience with cloud platforms and containers, such as AWS, Azure, or GCP, and Docker or Kubernetes.
Experience with agent evaluation frameworks and LLM or agent observability tooling.
Familiarity with the Model Context Protocol (MCP) or similar tool-integration standards.
Healthcare, insurance, or other regulated-industry background.
Experience with SQL and large-scale data processing, such as Spark.
Experience with MLOps practices, including feature stores, model registries, and automated retraining.

Responsibilities

Design and build multi-step, tool-using agentic AI systems that plan, call functions and APIs, retrieve context, and act within guardrails on complex claims and clinical workflows.
Build and optimize the orchestration layer, including tool and function calling, multi-agent coordination, state management, and external-system integration through protocols such as MCP.
Develop retrieval-augmented generation (RAG) pipelines over clinical and policy documents, with evaluation harnesses to measure agent accuracy and tool-use correctness and guard against regressions.
Take agents and models from prototype to production, including data pipelines, APIs, serving infrastructure, deployment, and monitoring.
Define and track evaluation metrics for agentic systems, including task success rate, tool-call accuracy, latency, cost per run, and hallucination rate, and build test suites to maintain quality as systems evolve.
Design human-in-the-loop patterns and guardrails so agents operate safely and auditably in a regulated environment.
Partner with clinical subject matter experts, product, and compliance to translate domain expertise and regulatory requirements, including CMS, HIPAA, and 42 CFR, into reliable systems.
Contribute to responsible-AI and governance practices, including documentation, bias evaluation, and oversight design for autonomous workflows.