Backend & Agent Engineer AI/Agentic Automation

We are looking for a hands-on Backend & Agent Engineer to design and develop scalable backend services and AI/agentic workflows for Medicaid R2R processes. The role combines strong Python/API engineering with agent orchestration, enterprise integrations, workflow state management, observability, security, and production deployment., * Develop scalable Python backend services, REST APIs, and integration components.

• Build and orchestrate AI agents for reconciliation, exception investigation, variance analysis, and close processes.
• Implement agent state, routing, tool/function calling, retries, checkpoints, and failure recovery.
• Integrate agents with enterprise data, SQL, APIs, rules engines, and governed tools.
• Implement human-in-the-loop review and approval workflows.
• Develop prompts, grounded context, structured outputs, and evidence-backed agent responses.
• Build automated unit, API, integration, regression, and agent-evaluation tests.
• Implement logging, tracing, monitoring, security, and production-readiness controls.
• Support CI/CD, deployment, configuration, and operational troubleshooting.

• Strong Python and SQL
• FastAPI or equivalent Python API framework
• REST APIs and microservices
• AI/LLM and Agentic AI concepts
• Agent orchestration and state management
• Prompt engineering and tool/function calling
• RAG and enterprise data integration
• Workflow orchestration and exception handling
• Automated testing and CI/CD
• Observability, logging, telemetry, and tracing
• IAM/RBAC, secrets management, encryption, and secure API design

Preferred Technology Experience

• LangChain / LangGraph
• Azure OpenAI
• Databricks / Delta Lake
• MLflow tracing and evaluation
• Unity Catalog
• MCP (Model Context Protocol)
• Databricks Workflows / Model Serving
• Azure DevOps or GitHub Actions
• Docker / containerized deployment, A strong Python backend and AI/agent engineer capable of translating Medicaid finance workflows into secure, scalable, observable, and production-ready agentic solutions, with a clear understanding of when to use deterministic business rules versus LLM-driven reasoning.