Principal AI Product Engineer

Join Hi Marley, a fast-paced startup revolutionizing the insurance industry. As a Principal AI Product Engineer, you'll build the internal infrastructure and tools that enable seamless collaboration across departments. You'll design internal knowledge systems, decision-making surfaces, and agent-builder UIs, while managing a portfolio of small internal products. This role requires a unique blend of design taste and technical expertise, as you'll be both designing and coding. You'll work closely with the Technical Infrastructure team and be based in our Boston office 2-3 days a week. Enjoy an open vacation policy, ample learning opportunities, and a full benefits package.

• Construire l'infrastructure et les outils qui permettent à chaque département de Hi Marley de travailler ensemble de manière harmonieuse.
• Comprendre comment les données circulent dans l'entreprise et les exposer au bon niveau pour le bon public.
• Gérer un portefeuille de petits produits internes, en prenant en compte les demandes de l'ensemble de l'entreprise, en priorisant par levier et en séquençant le travail.
• Product instinct, not just craft: You can hold ten internal asks in your head, prioritize by leverage, and ship a coherent portfolio of small, well-shaped products. Comfort with ambiguity and with saying no
• Taste as a moat: You make UI judgment calls beyond what's in the spec — spacing, motion, information density — because you've internalized why those things matter
• Bonus — AWS familiarity: Hi Marley is AWS-only (Cognito, Bedrock, CDK). You don't need to be deep here on day one — your infra counterpart owns the spine — but be willing to go deep when needed
• Systems thinking about data and tools: You think in terms of where data lives, what to centralize vs. decentralize, and how to expose it at the right level. You design with the data architecture in mind, not just the screen
• Modern frontend stack: Strong with React, Next.js, TypeScript, Tailwind, shadcn/ui (or equivalent). Comfortable touching the backend when the boundary is unclear
• Communication and craft: You can present your work to engineers, executives, and customers, and explain why you made a decision in language that matches the audience
• Lives in agentic coding tools: Claude Code, Codex, Cursor — these are your default environment. You don't just "use AI to help"; your craft is shaped by these tools
• Hybrid design + engineering chops: You have strong design instincts and you ship production code. You sketch, code, and merge your own PRs — no handoff