AI Team Lead

Summary

You will be the tactical engine of a growing AI team, shipping production agents alongside your engineers, unblocking the hard problems, and raising the bar for everyone around you. This is a player coach role: you'll be deep in the code, mentor the next generation of AI engineers, and make sure the squad delivers real business value sprint after sprint. Leadership here means pull requests and solved problems, not slide decks.

About Jewelers Mutual

Jewelers Mutual Group has been protecting the jewelry industry for over a hundred years, and we're now pairing that deep domain expertise with a ground-up technology transformation. Our engineering team in Raleigh is building a serverless, event-driven platform on AWS, a modern data stack on Databricks, and an enterprise AI capability designed to change how insurance works. This is Silicon Valley energy inside a company that's been earning trust since 1913.

What You'll Do

• Drive Execution on the AI Squad: Own the day to day technical delivery of the AI team. You'll set sprint priorities, break down complex work into shippable increments, remove blockers, and make sure the squad is consistently delivering value, not spinning on interesting problems that don't move the needle.
• Ship Code and Ship Agents: This is not a clipboard role. You'll be hands-on in the codebase, designing multi-agent orchestration systems, building integrations, and solving the hardest technical problems on the squad. You lead by example, and your team sees your commits.
• Lead and Grow AI Engineers: Manage a team of AI Engineers, providing technical mentorship, meaningful code reviews, and career development. You'll create an environment where builders thrive: high standards, fast feedback loops, and a bias toward shipping.
• Own Production Reliability for AI Systems: Establish the evaluation frameworks, observability practices, and operational standards that keep agents performing in production. When something breaks at 2 AM, your architecture and your runbooks should make the fix obvious.
• Drive AI Adoption Across the Business: Partner with leaders across underwriting, claims, customer experience, operations, and engineering to identify the highest-leverage opportunities for AI. Translate business problems into technical approaches and help stakeholders understand what's possible, what's practical, and what's next.
• Build AI That Thrives in Production: Anyone can build a compelling demo. You'll hold the team to a higher standard, building agents with the transparency, explainability, and ethical foundations that unlock deployment in workflows where it actually matters.
• Stay Ahead of the Curve: Maintain deep fluency in the rapidly evolving agentic ecosystem, including Claude, Claude Code, MCP, and emerging frameworks. Evaluate new tools and approaches with a builder's eye and make pragmatic bets on what to adopt, what to watch, and what to skip.

What You'll Bring

• 5-8+ years of software engineering experience, with 2-3+ years building and deploying AI/LLM-powered systems in production. You've shipped agents that handle real workloads and you've learned what breaks when they do.
• Prior experience leading a technical team, whether as a formal tech lead, engineering manager, or senior IC who naturally pulled others into their orbit. You know how to set direction, give hard feedback, and develop talent.
• Strong software engineering skills across one or more modern languages, with deep experience in agentic frameworks, including Lang Graph, CrewAI, Auto Gen, Claude tool use, or equivalents. You have opinions about these tools because you've built real things with them.
• Strong fundamentals in API design, async/event-driven systems, and cloud infrastructure (AWS strongly preferred). You can design systems that other engineers want to build on.
• Experience building evaluation, tracing, and observability for AI systems, using Lang Smith, Langfuse, or similar tools. You've defined what "good" looks like for agent behavior and built the systems to measure it.
• Working knowledge of RAG architectures, vector stores, and embedding pipelines. You've built retrieval systems on real-world, messy enterprise…