AI Engineer

Role & responsibilities

Junior Role Exp-3 to 6 yrs

Senior Role Exp  8 to 10 yrs

Overview:

Were looking for a hands-on engineer to help us build and tune intelligent agents designed as LangGraph-based graphs that also leverage MCP tools. These agents run on Azure and are central to our AI-driven automation for the insurance industry. The ideal candidate is comfortable designing, implementing, and debugging multi-step agent workflows in a cloud-native environment.

They should be familiar with the broader retrieval-augmented generation (RAG) stack including embeddings, vector databases, and chunking strategies as this will be helpful for scaling intelligent document understanding across submissions and claims.

Key Requirements:

• Strong experience with LangChain, LangGraph, and LangSmith
• Familiarity with MCP Tools and architecture 
• Experience with Azure AI Foundry and Functions and cloud-native deployment patterns
• Understanding of RAG, embeddings, and vector stores
• Comfortable with prompt engineering, agent orchestration, and debugging
• Bonus: domain knowledge in insurance or experience building enterprise-grade LLM applications

High-Level Product Vision:

Brisc is building a new class of AI-native infrastructure for the insurance sector — specifically MGAs and reinsurers. Our goal is to streamline high-friction, high-volume processes using intelligent agents that handle real-time decisioning, document ingestion, triage, and data reconciliation.

Current Agent Use Cases:

•  Submissions Agent – parses submission emails and documents, extracts data, and applying underwriting processing and knowledge, help priortize. The agent must be fully transparent and repeatable. 
•  Bordereaux Reconciliation Agent – automates premium and policy data matching across bordereaux files and internal systems.
•  Claims Notification Agent – ingests claim notices and surfaces critical items requiring human intervention
•  Supervisor Agent – coordinates graph-based task execution and ensures secure data handling per client
•  Bordereaux Extraction Agent – working with unstructured data, vision processing and LLMs for extraction. 

- 

We’ve already built initial versions of these agents and have a functional dashboard, database, and event infrastructure in place. The next step is refining accuracy, scaling deployment, and extending our agent capabilities as we prepare for production rollout and Azure Marketplace listing.