Specialty insurance is one of the largest and least digitised markets in the world. Every day, brokers and MGAs trade risk through application forms, email chains and manual re-keying. The same submission is typed out and extracted again at every step.

At Braid, we're changing that. Brokers can now transact with MGAs through a completely agentic channel, without the need to adopt another system, and receive instant quotes back from their panel. MGAs get the clean, structured submissions that fed straight into their own pricing and underwriting rules.

What you will do

As a Founding Engineer, you'll be working out how AI agents take part in a regulated market. That means:

- Building agent-ready systems. Design the MCP tools and multi-agent orchestration that let brokers' AI assistants discover products, qualify risks and compare quotes.
- Turning documents into data. Build the extraction pipeline that turns messy broker submissions (PDFs, spreadsheets, emails) into structured, verifiable risk data.




- Encoding underwriting logic. Turn MGA pricing and underwriting rules into decision graphs that give the same, explainable answer every time.
- Keeping it reliable. Own evals, audit trails and model portability, so the system can be trusted and can run on the cheapest model that does the job well.
- Owning the product with the founders. Work directly with the founders across the whole product and take ownership of large parts of the architecture and development process.

What we're looking for

- Strong experience with modern AI coding tools (such as Claude Code, Codex or Cursor).
- Strong experience building with LLMs, such as agent frameworks, MCP, tool use and evals.
- Python and TypeScript preferred.
- Robust written and verbal communication in English.
- Experience in insurance, fintech or agentic systems.

📌 Founding Engineer (India)
🏢 Braid
📍 India