Optum is a global organization that delivers care, aided by technology to help millions of people live healthier lives. The work you do with our team will directly improve health outcomes by connecting people with the care, pharmacy benefits, data and resources they need to feel their best. Here, you will find a culture guided by inclusion, talented peers, comprehensive benefits and career development opportunities. Come make an impact on the communities we serve as you help us advance health optimization on a global scale. Join us to start Caring. Connecting. Growing together.
Position in this function builds, configures and supports secure AI integrations that connect business-unit data and services to AI platforms. On Optum Technology's AI Engineering connector team, you will deliver plugins, connectors, Model Context Protocol (MCP) servers and related components across Microsoft Copilot, OpenAI, Anthropic and Google ecosystems. This is a hands-on implementation, testing and operational support role, not primarily model training or ETL development. Partner with source owners, platform teams and reviewers to move approved integrations from intake through validation, launch and ongoing support.
Primary Responsibilities:
•Translate business intake into implementation designs with source owners and AI Engineering; confirm approved scope, data boundaries, acceptance criteria and leadership priorities
•Build and configure AI connectors, plugins, MCP servers and, where approved, skills packages. Validate submitted manifests, API contracts, tool schemas, skill and instruction files, platform-specific packaging and custom code before onboarding; account for instruction differences across platforms
•Implement approved authentication and authorization, including Entra app registrations, delegated OAuth, least-privilege permissions and source-to-platform access mapping; coordinate IAM approvals
•Deliver safe retrieval and, where explicitly approved, write capabilities that respect source entitlements, data-use limits and enterprise gateway requirements
•Engineer reliable API behavior with pagination, bounded retries, rate-limit handling, synchronization controls and actionable error handling
•Create unit, integration, end-to-end, security and performance tests; validate authentication, retrieval, access mapping and sync behavior in sandbox and staging environments
•Use formal intake with technical evidence for architecture, Security, Legal and privacy reviews, including AIRB/ESRO where applicable. Retain catalog provenance through AID or approval metadata where required to distinguish reviewed artifacts from unauthorized uploads. Implement approved controls without substituting for approval owners; help control owners define evolving custom-agent governance
•Administer platform configuration and registrations, including M365 where relevant, and explain integration usage. Deploy only through approved development, test and production environments, least-privilege access and release gates; prevent direct production creation or uploads that bypass review. Maintain reusable configuration and onboarding templates
•Maintain version-controlled code, peer reviews and CI/CD checks; document API changes, configuration, deployment and recovery procedures
•Instrument monitoring and alerts, protect sensitive data in logs and maintain runbooks for support, access changes, onboarding and audit evidence
•Diagnose integration failures and network reachability, internet, gateway and firewall dependencies with AI Catalog, cloud, platform, network and source teams. Revert configuration or roll back releases safely, establish root causes and preventive fixes, and escalate to vendor technical support with reproducible evidence
•Supply test evidence and support readiness for pilot, limited availability (LA) and general availability (GA) gates; work with the manager and project manager on risks, dependencies and user feedback
•Comply with all applicable Company policies, procedures, and business directives, changes including those relating to work location, team assignments, work schedules, and flexible work arrangements
Required Qualifications:
•Primary: demonstrated AI integration development - Demonstrated hands-on development, deployment, platform configuration and troubleshooting of connectors, plugins or MCP servers through a validated pilot or production release on at least one AI platform, with transferable diagnostic skills across ecosystems. Explain architecture, package structures, tests and failure handling; diagnose authorization/token endpoints, service principals, client IDs and callbacks, including differences between Entra and source-managed or custom identity providers. Generic model-training, ETL or UI-only administration experience alone does not meet this requirement
•Experience connecting enterprise APIs or data sources to at least one AI platform, with practical understanding of authentication, source authorization mapping and safe retrieval or approved tool execution
•Secondary: programming and software delivery - Bachelor's degree in computer science, engineering or a related field, or equivalent practical experience. Demonstrated problem-solving and clear technical documentation for implementation and support partners
•Solid Python or Java development experience and ability to work with the other language. Three or more years of hands-on software or integration development, or equivalent demonstrated capability, proposed for HR review. Demonstrate reading and debugging custom integration code, validating submitted packages, and delivering tested configuration and deployment changes. Working knowledge of REST/JSON, OAuth, automated testing, version control, peer review and CI/CD
•Tertiary: governance, legal review and data privacy - Familiarity with AI and enterprise governance, legal review, data privacy and least-privilege principles. Ability to document data flows and controls, participate in required reviews, implement approved restrictions and escalate risks. Approval authority remains with designated architecture, Security, Legal, privacy and IAM owners
