Principal Product Manager - Technical, Agentic AI Core Infra

Amazon Amazon · Big Tech · NY +1 · Project/Program/Product Management--Technical

Principal Technical Product Manager to lead the strategy and execution of an end-to-end agentic core infrastructure (SDK, Runtime, Builder Tools) that enables agent and skill builders to develop, test, and operate production-grade agentic experiences at scale. This role will innovate leveraging and extending AWS core primitives like Agent Core to fulfill Amazon Ads specific needs like agent authorization, guardrails, shared memory, and fully managed runtime for complex multi-step reasoning and multi-agent workflows.

What you'd actually do

  1. Own the vision and strategy for building Agentic AI Core Infra products and features
  2. Own the vision and strategy for accelerating agentic transformation Ads-wide
  3. Write documents that work backwards from different types of customers, including advertisers and agent builders
  4. Dive deep into competitive and existing Amazon offerings
  5. Partner effectively with other leaders in product, design, engineering, science and business intelligence to influence priorities and drive alignment

Skills

Required

  • 2+ years of end to end product delivery experience
  • 8+ years of technical product or program management experience
  • Bachelor's degree
  • Experience with feature delivery and tradeoffs of a product
  • Experience owning/driving roadmap strategy and definition
  • Experience leading engineering discussions around technology decisions and strategy related to a product
  • Experience technical product management

Nice to have

  • Experience leading cross-organizational programs and setting and holding teams and individuals to timelines and deliverables without explicit authority
  • Experience working with and influencing senior level stakeholders

What the JD emphasized

  • agentic core infrastructure
  • agent and skill builders
  • production-grade agentic experiences at scale
  • multi-step reasoning
  • multi-agent workflows
  • agent authorization
  • guardrails
  • shared memory
  • fully managed runtime

Other signals

  • building agentic core infrastructure
  • enabling agent and skill builders
  • production-grade agentic experiences at scale
  • multi-step reasoning and multi-agent workflows