AI Technical Lead (Principal Engineer)

As our Technical Lead, you will design and build a new kind of product: an insurance marketplace that AI agents use to shop on a person’s behalf. You will own the technical architecture end to end, from the models and how they run to the MCP server and tools that other AI platforms connect to. This is a hands-on senior role on a small team, and the decisions you make early will shape what agentic insurance shopping looks like before the market arrives.

You own the technical architecture for the agent: where the model’s reasoning runs, how we tune and serve it, how we keep cost and speed under control, and how the whole system stays reliable. We intend to run and tune our own open-weight models rather than depend only on hosted APIs, so we control cost, speed, and behavior as volume grows, and owning that is central to this role. You own the MCP server, the tools, and the data the agent works with, and you set the patterns the rest of the team builds on. You are the person who decides, with testing to back it up, how this is built, and then builds it.

- Own the architecture: decide where the agent’s reasoning runs, which models we use, how they are packaged, and how AI platforms connect to us, and back those calls with testing.

- Build and tune the models: fine-tune open-weight models for our use, and decide when tuning, retrieval, or prompting is the right tool.

- Stand up the serving and infrastructure: deploy the agent and MCP server so they are reliable, fast, and affordable to run, with versioning and a way to roll back.

- Build the testing into the system: programmatic eval sets, guardrails, and monitoring for quality, drift, cost, and speed, so we catch problems before agents do.

- Design the data model and tools with the AI product lead: the request and response payloads, the reasoning we expose, and the trust signals.

- Set the engineering patterns and standards the team builds on, and keep the code and architecture documented.

- Keep cost and speed honest: watch what inference costs, and make the trade-offs that keep it sustainable as volume grows.

- Work in the open. Help publish our code, architecture, and findings.

- Mentor the engineers on the team and give leadership clear, regular visibility into technical progress and risks.