AI Product Manager

Shepherd • San Francisco, CA

8d • $190,000 - $220,000 • Onsite

About The Position

Shepherd is a technology-driven Managing General Underwriter (MGU) transforming commercial Property & Casualty insurance for high-hazard industries. Our mission is to make risk frictionless for the builders and operators shaping the physical world, protecting progress from concept through construction and into decades of operation. We're building the fastest, smartest commercial risk platform, where underwriting expertise, data, and automation work together to deliver faster decisions, smarter, more accurate pricing, and better risk outcomes. With Shepherd, safety, speed, and quality no longer trade off against one another. They compound. We're not just modernizing insurance products. We're building the risk infrastructure for the next generation of financial services, where technology, underwriting, and partnerships operate in harmony to support the world's most important industries and the progress they make possible.

This role is for the Nova Team, which is building fully autonomous underwriting. The team is small, extremely high caliber, and growing. The engineers have taken autonomous underwriting from an idea to real prices on real accounts, making the product calls themselves. This role sits alongside leadership and at the intersection of three domain expertises: engineers who think in eval runs and retrieval quality, underwriters with decades of pricing judgment, and leadership setting the direction of the business. The business case is straightforward: our long term vision has a single underwriter overseeing roughly 10x the accounts they handle today, with the system doing intake, analysis, and pricing while the person sets strategy and works exceptions. Getting there is largely a measurement challenge. An account only counts as autonomous if everything on it is right, which makes this a compounding problem rather than an average one, and much of the roadmap comes down to how we define and measure that.

Requirements

• 4+ years in product, having owned the direction of a hard technical product and can point to what came of it.
• Technical in practice: Can read a PR, contribute in an architecture discussion, write your own SQL, and reason about why a model behaves the way it does.
• Shipped AI into production: Knows the difference between a demo and a system people rely on, and has informed opinions about evals and model quality.
• Statistically literate: Comfortable reasoning about distributions and thresholds, and interrogates a metric before trusting it.
• Willing to go deep on the domain: Understanding construction risk well enough to encode it is most of the advantage here.
• A direct communicator: Writes clearly, disagrees in the room, and doesn't need a deck to make a decision.
• Opinionated, high agency, experimental: Forms a point of view and argues for it. Doesn't wait to be handed work. Instinct when something is unclear is to run the experiment rather than debate it.
• No playbook and no product org to lean on here, so you'll define how product works and change it as the team grows.

Nice To Haves

• Background in autonomy, robotics, or another field where escalation and graduated trust are core concepts
• Document AI, information extraction, or agentic workflow experience
• Time in insurance, fintech, or another regulated industry
• Founder or early-employee experience
• Hands-on with eval tooling and experiment tracking

Responsibilities

• Own the product end to end: Vision, discovery, prioritization, execution, and the UX and design decisions along the way.
• Prioritize ruthlessly, be as clear about what we're not building as what we are, and shape how we staff design and engineering against the roadmap.
• Be accountable for results rather than shipped features: adoption by our underwriters, how much of the work runs without a person, how accurate it is, and whether we can win a deal on a fully autonomous motion.
• Understand the user deeply: Sit with underwriters, read submissions yourself, and build the instinct for why one contractor's paperwork is far harder to handle than another's.
• Find the places where the system can make their work faster, more reliable, and more complete, and turn that into requirements and edge case coverage.
• Own the definition of correct: Shape what gets measured, where materiality thresholds sit, which regressions block a release, and how field-level results roll up into claims we're willing to make publicly.
• Set the bar for what autonomous means at Shepherd: How our agents reason, when they act on their own, when they defer, and how they earn an underwriter's trust.
• Own how uncertainty gets surfaced and what a person sees when the system hands something back.
• Ship fast with engineering: Write prompts, dig through eval data, and pull the failing runs to work out whether a problem is retrieval, model behavior, or how we've modeled the data, all before filing a ticket.
• Use AI coding agents to prototype, validate, and unblock, and write code yourself when that's the quickest path, without letting the quality bar slip.
• Think big, build big: Aim for a system that runs on its own, and sequence the roadmap backward from that rather than forward from what's easy.
• Win together: Support each other, raise the bar, and celebrate collective success. As the first PM, set a standard the rest of the team inherits, and ensure milestones belong to the team rather than to product.
• Cross the aisle: Listen deeply, work across boundaries, and prioritize shared success over individual lanes. Make product calls that come from engineers who've sat with underwriters and underwriters who understand where the model breaks.
• Go get it: Act with urgency, move with confidence, take smart risks, and push forward with intention. Pull the failing runs, book the time with the underwriters, and decide what matters.

Benefits

• Premium Healthcare: 100% contribution to top-tier health, dental, and vision
• Fertility benefits and family building support
• Unlimited PTO
• Flexibility to take the time off, recharge, and perform
• Daily lunches, dinners, and snacks
• SF, NYC, Dallas-Fort Worth, Chicago and LA Offices
• Professional Development: Access to premium coaching, including leadership development
• Competitive 401(k) Plan
• Dog-friendly office

What This Job Offers

• Job Type: Full-time
• Career Level: Entry Level
• Education Level: No Education Listed