Role Description:

AI Builders is ShipBob’s new job family — the deliberate next step for Product Management in the AI-native era. The thesis is simple: when AI tools can generate working code from precise specs, the slowest step is no longer the spec — it is the PM→Engineer handoff. AI Builders close that gap by owning discovery, specs, prototypes, and production PRs themselves, accelerated by an AI-native pod.

As a Senior AI Product Builder at ShipBob, you own the full build loop for your product domain: discovery, story/specs, AI-augmented prototypes, and production-ready PRs. This is not the traditional PM role — you are a builder. You use the AI development toolchain (e.g., Claude Code, Driver.ai) to understand ShipBob’s codebase architecture, generate working code against the actual repo, and ship with minimum engineering rework. The minimum technical floor for this role is real: you must read ShipBob’s data model, evaluate AI-generated scaffolding for correctness, and write story/specs precise enough that AI tools can act on them without follow-up questions.

Your domain is Inventory Optimization — the decision systems that determine where a merchant's inventory should live across ShipBob's fulfillment network, how much should be positioned at each node, and when to replenish before stockouts or overstock erode margin. IO sits at the intersection of demand forecasting, inventory placement, Promise, and the broader Merchant Tech surface. Placement quality is upstream of delivery speed, fulfillment cost, and the promises ShipBob can make to shoppers — it is one of the highest-leverage product surfaces in the company.

This role is for someone energized by speed, uncomfortable with waste, and who measures success by outcomes shipped — not documents approved. You operate in an AI-native pod alongside talented engineers who are ready to support you and push back on solution when needed. This role reports to a Director, AI Product Builder.

What you'll do:

Own the full product loop end-to-end: AI-augmented discovery → story/spec → prototype → production PR → outcome measurement. You are accountable for the outcome, not just the handoff.
Own the forecasting, placement, and replenishment decision surface end-to-end. Define the product contracts for demand forecasts and distribution recommendations, ship them into merchant workflows, and measure adoption and forecast accuracy.
Run customer discovery yourself; use AI tools to synthesize research across tickets, interviews, behavioral data, and competitive signals.
Demonstrate deep functional and technical knowledge of owned products end-to-end. Demo to merchants and partners yourself; understand competitive landscape and key differentiators.
Use AI tools to analyze product metrics, identify usage patterns, and surface anomalies. Define success metrics and kill thresholds at spec time — without coaching.
Maintain a prioritized roadmap ranked by impact × speed. Use AI tools to synthesize competitive intelligence, market trends, and customer insight into strategic inputs.
Evaluate whether a proposed solution fits ShipBob’s architecture without Engineering translation. Understand service boundaries, data structures, and how merchant and partner API usage shapes data contracts.
Route changes through prototype-then-PR. Write clear story/specs and review AI-generated code before raising. PR promotion rate improves consistently.
Set experiment hypotheses with explicit kill thresholds. Make day-7 rollout/kill decisions on AI-surfaced data.
Participate in architecture reviews alongside Engineering as the product judgment seat — the person making the build-vs-buy-vs-defer call on the room’s behalf.
Additional duties and responsibilities as necessary.

Your first 90 days:

Days 1–30 — Map the domain. Understand the ShipBob operating and commercial mode, examine the IO codebase, anchor on the purpose and value of the domain, and stakeholder conversations to triangulate the top three active bets. Demo a throwaway prototype back to the pod.
Days 30–60 — Ship your first PR. Small and safe is fine. The point is proving the loop, not the size of the change.
Days 60–90 — Make a real call. Define and instrument the metric for one real bet, with the kill threshold set at spec time. Make at least one rollout-or-kill call on the data, in front of the team.

What you'll bring to the table:

5+ years of product management, product engineering, or equivalent experience, with demonstrated hands-on AI tool usage in a professional context.
Demonstrated ability to write story/specs that AI tools can act on directly — state machines, business rules, data contracts, edge cases — not narrative PRDs.
Technical fluency: can read a data model, evaluate an API contract, and identify when AI-generated code deviates from the target architecture.
Track record of shipping working prototypes or production code, not just Figma mocks or slide decks.
Sound outcome ownership: sets success metrics and kill thresholds at spec time and tracks them after launch.
Experience working in supply chain, fulfillment, logistics, transportation, or e-commerce technology.
Comfortable operating in a fail-fast culture. Being wrong fast is a feature here — staying slow to look right is the failure mode.
Excellent written communication: you write specs, not meeting notes.

Nice to have:

Hands-on experience with the AI development toolchain (e.g., Claude Code, Driver.ai, Cursor, Replit, V0, Figma AI).
Working knowledge of Python, SQL, or TypeScript — enough to debug AI-generated code and interrogate the data behind a forecast.
Experience with inventory planning, demand forecasting, replenishment, or multi-node placement and distribution optimization at scale.
Prior experience transitioning a product team from traditional SDLC to an AI-native operating model.