Join Optum, a global leader in healthcare technology, and contribute to improving health outcomes for millions. Our team is dedicated to connecting people with the care, resources, and data they need. You'll find an inclusive culture, collaborate with talented colleagues, and benefit from comprehensive career development.

This role focuses on advancing health optimization globally. We are committed to Caring, Connecting, and Growing together to make a significant impact on the communities we serve.

Develop and implement agentic workflows utilizing Semantic Kernel, focusing on tool calling, multi-agent coordination, and safe tool execution. Design and build robust Retrieval Augmented Generation (RAG) pipelines, incorporating chunking, embeddings, retrieval, and ranking with various vector stores like Azure AI Search, Pinecone, or Weaviate.

Deploy scalable services using C# or Typescript within Docker and Kubernetes environments, ensuring adherence to CI/CD practices, observability standards, and optimization for cost and latency. Integrate with enterprise systems through APIs, message buses, and event-driven architectures, while also designing appropriate function and tool schemas.

Apply Responsible AI principles, including prompt injection defenses, content filtering, PII/PHI redaction, audit logging, and human-in-the-loop mechanisms. Establish and maintain offline and online LLM evaluations for accuracy, hallucination rates, and safety using frameworks like Ragas or DeepEval. Mentor junior developers to foster team growth and knowledge sharing.

Possess a Bachelor's degree in Computer Science, Engineering, or a related field. Demonstrate at least 7 years of professional experience, with strong proficiency in programming languages such as C#, Java, Python, or Go.

Experience with cloud-native solutions on Azure, AWS, or GCP is essential. Familiarity with enterprise system integration via APIs, message buses, and event-driven architectures is required. A solid understanding of Responsible AI practices and LLM evaluation frameworks is also necessary.