Prototyping Architect (physical Ai), Aws Prototyping and AI Customer Engineering (pace)

Amazon Amazon · Big Tech · Seattle, WA · Solutions Architect

Builds prototypes for customers at the intersection of Generative AI, Physical AI, and robotics, leveraging AWS services and NVIDIA technologies. Focuses on architecting and developing solutions involving digital twins, simulation, synthetic data, and robot training to solve real-world industrial and autonomous system problems.

What you'd actually do

  1. Architect and build working Generative AI, Agentic AI, and Physical AI prototypes directly with customers using AWS AI services (Bedrock, SageMaker, IoT TwinMaker, IoT SiteWise) and cloud-native architectures - including autonomous agents, RAG architectures, LLM-powered applications, and simulation-driven workflows that demonstrate production-ready solutions
  2. Design and build Physical AI prototypes spanning digital twin environments, discrete event simulation, robotic policy training pipelines, and synthetic data generation - leveraging tools such as NVIDIA Omniverse, Isaac Sim, AWS VAMS, and AWS-native compute and storage services
  3. Leverage AI-driven development tools (Cursor, Kiro, Q Developer, Claude Code) to accelerate prototype development, implementing patterns like prompt engineering, function calling, agent orchestration, tool use, and simulation pipeline automation
  4. Serve as a trusted technical advisor to customers on LLM selection, agent design, Physical AI architecture, and AI adoption strategy - guiding them through complex technical decisions and trade-offs across both software-defined AI and physical systems
  5. Collaborate with Technical Program Managers, Design Technologists, and fellow Prototyping Architects to deliver customer engagements on time and with lasting impact, working across the full Physical AI flywheel from spatial data and simulation through to model training and deployment

Skills

Required

  • Generative AI
  • Agentic AI
  • Physical AI
  • robotics
  • simulation
  • digital twins
  • synthetic data generation
  • AWS AI services (Bedrock, SageMaker, IoT TwinMaker, IoT SiteWise)
  • NVIDIA Omniverse
  • NVIDIA Isaac Sim
  • AWS VAMS
  • full-stack development
  • prompt engineering
  • function calling
  • agent orchestration
  • tool use
  • simulation pipeline automation

Nice to have

  • AWS-native compute and storage services
  • AI-driven development tools (Cursor, Kiro, Q Developer, Claude Code)
  • LLM selection
  • Physical AI architecture
  • AI adoption strategy
  • spatial data
  • model training
  • deployment

What the JD emphasized

  • Physical AI
  • prototypes
  • simulation
  • robots
  • digital twins
  • synthetic data generation
  • robotic policy training
  • agent orchestration
  • tool use
  • simulation pipeline automation

Other signals

  • prototypes
  • physical systems
  • simulation
  • robots
  • digital twins
  • synthetic data