Senior Product Manager, Tech, Amazon Manufacturing Services

Amazon Manufacturing Services (AMS) is seeking a Sr. Product Manager, Tech to own the product vision, strategy, and roadmap for the manufacturing execution and automation intelligence stack powering Amazon's first advanced manufacturing facility — a highly automated, first-of-its-kind operation. This facility integrates industrial robotics, end-to-end manufacturing automation, and digital manufacturing to produce systems for Amazon's global fulfillment network. Key job responsibilitiesManufacturing Execution Product StrategyOwn the 2–3 year product roadmap for the manufacturing execution stack — from scheduling UX through operator tools to AI-powered optimizationDefine the product vision that connects AI-native scheduling intelligence to the physical shop floor experienceDevelop and maintain the execution technology strategy across P0 (manual + semi-automated), P1 (AGV integration, real-time optimization), and P2 (AI-enabled autonomous control) Write compelling narratives (PR/FAQs, 6-pagers, OP docs) that articulate product strategy and secure investment from leadershipScheduling & OperatorExperienceServe as the operational product partner to the scheduling engineering team — translating manufacturing floor needs, pain points, and workflows into product requirementsOwn the operator-facing scheduling experience: how assignments surface, how disruptions are communicated, how overrides are captured, and how the system explains its decisions to the right audienceDrive the scheduling system's authority progression (Shadow Advisory Co-Pilot Decision Maker) by defining success criteria, measuring override rates, and building supervisor trust through UX designDefine product requirements for scheduling and explainability tools from the operator/supervisor perspectiveMachine Connectivity & IIoTOwn the product vision for machine connectivity — defining how equipment telemetry flows from factory floor to data lake to decision systemsDefine integration requirements for manufacturing equipment onboarded at the manufacturing facility (industrial lasers, robotic welding cells, automated coating lines, autonomous mobile robots) Develop product requirements for digital twin capabilities — real-time equipment state, simulation for what-if analysis, and predictive modelingPartner with automation engineers to define the machine-to-cloud data contract for each equipment classAI/ML & Continuous ImprovementDefine and execute product strategy for AI-driven manufacturing intelligence: predictive maintenance, automated quality inspection, process parameter optimization, and autonomous cell controlOwn requirements for the data platform layer that enables ML — feature stores, event streams, model serving infrastructureDrive the feedback loop between quality outcomes (first-pass yield, scrap rates) and upstream process adjustments — ensuring the system learns and improves continuouslyDefine the operator interaction model for AI recommendations — when to alert, when to auto-act, when to require human confirmationShop Floor Quality & ComplianceOwn the product experience for in-line quality: inspection workflows, non-conformance reporting, root cause analysis tools, and SPC dashboardsDefine how quality data flows back to both the scheduling system (for replan triggers) and the enterprise system (for financial variance reporting) Partner with Quality Engineering to translate ISO 9001:2015 requirements into tool capabilities and audit-ready data records A day in the lifeIn the morning, you review overnight production data from our prototyping factory — the scheduling system's override rate dropped to 7% this week, and you're preparing the case to promote it from Shadow to Advisory mode. You pull the override-reason breakdown to identify two UX issues driving unnecessary supervisor interventions, then draft requirements for the engineering team. Mid-morning, you join the scaled factory equipment onboarding review. The laser cutting integration team needs a decision on telemetry frequency — you work through the tradeoffs between data granularity (better for predictive maintenance models) and network cost (lower in batch mode), landing on a tiered approach by signal type. After lunch, you're on the prototype factory floor shadowing a powder coating operator through a disruption scenario — a rush order just reshuffled the queue, and you observe how the shop floor app communicates the change. The operator missed the notification; you sketch a design change in your notebook. You close the day reviewing the ECO blast-radius prototype with the engineering team. An engineering change landed that affects 47 in-flight orders — the impact visualization needs to surface at-risk-dollars more prominently for the Change Control Board's decision meeting tomorrow.