AI Product Manager
$190,000–$220,000 year
On-siteSan Francisco, California, United States
Job Summary
Own the end-to-end roadmap for Nova, Shepherd's autonomous underwriting platform, by defining vision, discovery, and prioritization while ruthlessly shaping engineering and design staffing against adoption and accuracy metrics. Deepen understanding of user needs by sitting with underwriters, analyzing submissions, and translating friction points into requirements and edge case coverage. Establish the definition of correct by designing the eval framework, setting materiality thresholds, and determining which regressions block releases. Set the bar for autonomy by defining agent reasoning, escalation paths, and trust mechanisms, ensuring uncertainty is surfaced clearly. Ship fast with engineering by writing prompts, debugging eval data, prototyping with AI agents, and coding directly when necessary to unblock failing runs.
Required Qualifications
- 4+ years in product
- Track record owning the direction of a hard technical product
- Ability to point to what came of the built product
- Technical capability to read a PR
- Ability to contribute in an architecture discussion
- Ability to write SQL
- Ability to reason about why a model behaves the way it does
- Experience building things at work or on your own
- Experience shipping AI into production
- Knowledge of the difference between a demo and a system people rely on
- Informed opinions about evals and model quality
- Statistical literacy in reasoning about distributions and thresholds
- Ability to interrogate a metric before trusting it
- Willingness to go deep on the construction risk domain
- Ability to understand construction risk well enough to encode it
- Ability to write clearly
- Ability to disagree in the room
- Ability to make a decision without a deck
- Ability to form a point of view
- Ability to argue for a point of view
- High agency
- Experimental mindset
- Instinct to run an experiment rather than debate when something is unclear
- Ability to define how product works
- Ability to change product processes as the team grows
Desired Qualifications
- Background in autonomy, robotics, or another field where escalation and graduated trust are core concepts
- Document AI experience
- Information extraction experience
- Agentic workflow experience
- Time in insurance
- Time in fintech
- Time in another regulated industry
- Founder experience
- Early-employee experience
- Hands-on experience with eval tooling
- Hands-on experience with experiment tracking
Hiring someone like this?
Get your role in front of qualified candidates on Sorce.