Generative AI Architect - RAG/LLM

Key Responsibilities :

- Design and Implement Intelligent Agents : Lead the architecture, development, and deployment of sophisticated multi-step intelligent agents using LangGraph for complex workflows.

- Integrate and Optimize AI Tools : Leverage MCP tools effectively within agent designs to enhance functionality and performance.

- Cloud-Native Deployment : Implement and manage agent deployments on Azure AI Foundry and Functions, ensuring scalable, robust, and efficient operations.

- RAG Stack Optimization : Work with the broader retrieval-augmented generation (RAG) stack, including embeddings, vector databases, and chunking strategies, to enable intelligent document understanding across insurance submissions and claims.

- Agent Orchestration & Debugging : Comfortably engineer prompts, orchestrate agent interactions, and meticulously debug complex multi-step agent behaviors.

Develop Specific Agent Use Cases :

- Submissions Agent : Build agents to parse submission emails and documents, extract critical data, apply underwriting processing and knowledge, and prioritize tasks with full transparency and repeatability.

- Bordereaux Reconciliation Agent : Automate the matching of premium and policy data across bordereaux files and internal systems.

- Claims Notification Agent : Develop agents to ingest claim notices and surface critical items requiring human intervention.

- Supervisor Agent : Design and implement a supervisor agent to coordinate graph-based task execution and ensure secure data handling per client.

- Bordereaux Extraction Agent : Work with unstructured data, vision processing, and LLMs for accurate data extraction.

Required Skills & Experience :

- Expertise in LLM Frameworks : Strong hands-on experience with LangChain, LangGraph, and LangSmith.

- Cloud AI Platforms : Familiarity with Azure AI Foundry and Functions and best practices for cloud-native deployment patterns.

- RAG Fundamentals : Deep understanding of RAG architectures, embeddings, and vector stores.

- Prompt Engineering : Proven ability in prompt engineering, agent orchestration, and effective debugging of AI workflows.

- Architecture Acumen : Familiarity with MCP Tools and architecture.

Bonus Qualifications :

- Enterprise LLM Applications : Demonstrated experience building and deploying enterprise-grade LLM applications.