Principal Product Manager - Tech, AGI

The AGI Sensory & ML Builder Services is looking for a Principal Product Manager - Technical (L7) to help shape and develop the next generation infrastructure solution to train and build Amazon’s foundation models.
Amazon AGI is innovating rapidly to build and release AI Foundation models, and to support these initiatives the AGI Sensory & ML Builder Services team is building a unified modeling solution to enable experimenting and training across model modalities with minimum friction. This unified solution will be built on top of the our purpose-built high performance compute system to drive efficiency, high resiliency, and best in class performance for training some of the world’s largest models. The PMT will dive deep on the different foundation model building use cases and will work closely with science and MLE teams across AGI to capture requirements, define optimal user experience, and set roadmap for the engineering and product team to deliver on.

The ideal candidate will have outstanding communication skills, extensive product management experience, passion and familiarity with AGI, and ability to bring clarity in a highly ambiguous environment. They will be comfortable with complex architecture and will enjoy working in a fast-paced environment with high impact.


Key job responsibilities
• Build solutions to drive efficiency and lower builder friction, meet a high bar for UX, and enable innovation at scale
• Manage and execute large projects from start to finish including project management, defining core and guardrail metrics, problem solving, defining requirements and managing technical execution of changes.
• Engage with stakeholders and customers throughout the entire product lifecycle
• Define and drive creative, detailed product definition of new capabilities
• Manage prioritization and trade-offs among competing high value opportunities
• Work closely with engineering teams to deliver products to users
• Partner across organizations to maximize leverage and minimize duplication of work