Sr. AI/ML Engineer- Remote or Eden Prairie, MN

UnitedHealth Group•Eden Prairie, MN

9d•$120,100 - $214,500•Remote

About The Position

As a Sr AI/ML Engineer on the United AI Studio team, you will build agentic experiences on an enterprise AI platform—including Agents, Model Context Protocol (MCP) servers, plugins, and skills that enable teams to register, discover, and govern AI assets. You will build agents that eliminate friction across engineering workflows, starting with responsible AI reviews. You will join a team running AI Solution governance, AI Catalog, and enterprise-wide agent/skill registries, working alongside teams building underlying governance rails to deliver cutting-edge AI capabilities across the organization. You’ll enjoy the flexibility to work remotely from anywhere within the U.S. as you take on some tough challenges.

Requirements

Bachelor’s degree or 5+ years of equivalent software/AI engineering experience
5+ years of experience in backend software engineering, APIs, and service integration
2+ years of hands-on production experience with code-first agent frameworks (e.g., LangGraph, Google ADK, AWS Strands, Claude Agent SDK, OpenAI Agents SDK, Pydantic AI)
2+ years of experience building and deploying agentic solutions used in production environments, with a proven ability to debug and resolve production issues
1+ years of experience designing, building, or integrating Model Context Protocol (MCP) servers, tools, skills, or plugins
1+ years of experience implementing systematic evaluation harnesses or testing frameworks for LLM/agent behavior
Demonstrated experience using AI coding assistants (e.g., Claude Code, Codex, Copilot) as part of daily development workflows

Nice To Haves

Hands-on experience with managed agent runtimes (Azure AI Foundry Agent Service, Vertex AI Agent Engine, Bedrock AgentCore) and containerized deployments (GKE, EKS, AKS)
Experience with retrieval and semantic search (embeddings, vector stores, RAG patterns, chunking, reranking with Postgres/pgvector, Azure AI Search, or OpenSearch) and evaluating retrieval efficacy
Experience with LLM evaluation tooling (Ragas, DeepEval, Ranx, Braintrust, LangSmith, Langfuse, promptfoo) integrated directly into CI/CD pipelines
Experience routing model calls through enterprise gateways to manage failover, cost attribution, and model version pinning
Knowledge of durable execution for long-running agents (checkpointing, resumability, idempotent tool calls) and enterprise fundamentals (SSO, RBAC, audit, secrets management)

Responsibilities

Design, develop, and deploy AI-powered solutions to address complex business challenges with an emphasis on responsible use of AI
Build agentic experiences end-to-end, managing tool design, context and state management, failure handling, retries, and cost/latency budgets from prototype through production behind enterprise gateways
Design orchestration architectures, choosing between single agents, subagents, and deterministic pipelines, evaluating when a workflow should or should not be agentic
Expose AI capabilities through channel-agnostic APIs and Model Context Protocol (MCP) tools with required authentication, scoping, and auditability
Deliver agentic experiences into enterprise search and chat surfaces (e.g., Teams, Slack, Glean) using bot frameworks, connectors, and APIs
Stand up evaluation harnesses for agent behavior using golden task sets, trajectory/end-state scoring, LLM-as-judge calibrated against human labels, and regression suites integrated into CI/CD pipelines
Instrument production agents with tracing across tool calls, token/cost telemetry, and failure taxonomies, feeding real traces back into evaluation sets
Test agents adversarially against prompt injections and tool output vulnerabilities, implementing necessary remediations
Leverage enterprise-approved AI coding tools to streamline workflows, automate engineering tasks, and raise technical execution standards across teams