Member of Technical Staff - Hardware Science, Frontier AI & Robotics (far)

Amazon Amazon · Big Tech · San Francisco, CA · Machine Learning Science

This role focuses on foundational research and building intelligent robotic systems by developing foundation models for perception and manipulation, integrating them with hardware systems, and deploying them at Amazon scale. It involves independent research initiatives, full-stack robotics projects from conceptualization to hardware deployment, and collaboration with hardware engineering teams.

What you'd actually do

  1. Drive independent research initiatives across the full robotics stack, including robot co-design, manipulation mechanisms, innovative actuation and motor control strategies, state estimation, low-level control, system identification, reinforcement learning, and sim-to-real transfer, as well as foundation models for perception and manipulation
  2. Lead full-stack robotics projects from conceptualization through hardware deployment, taking a system-level approach that integrates actuator dynamics, sensor feedback (force/torque, IMUs, encoders), and electromechanical constraints with algorithmic development
  3. Develop and optimize control algorithms and sensing pipelines for physical robotic hardware, including motor characterization, actuator performance tuning, and robust sensor integration in production environments
  4. Collaborate with hardware, mechanical, and electrical engineering teams to ensure seamless integration of learned models across the robotics stack—from embedded compute and communication buses to actuator-level control
  5. Contribute to the team's technical strategy and help shape our approach to next-generation hardware-aware robotics challenges, including hardware-in-the-loop validation and prototype-to-deployment transitions

Skills

Required

  • PhD, or Master's degree and 4+ years of CS, CE, ML or related field experience
  • C++, Python, Java or Perl
  • hardware design
  • low-level control
  • state estimation
  • system identification
  • complex electrical systems
  • patents or publications at top-tier peer-reviewed conferences or journals
  • building machine learning models or developing algorithms for business application
  • applying theoretical models in an applied environment

Nice to have

  • reinforcement learning
  • sim-to-real transfer
  • foundation models for perception and manipulation
  • robot co-design
  • manipulation mechanisms
  • actuation and motor control strategies
  • actuator dynamics
  • sensor feedback (force/torque, IMUs, encoders)
  • electromechanical constraints
  • algorithmic development
  • sensing pipelines for physical robotic hardware
  • motor characterization
  • actuator performance tuning
  • robust sensor integration in production environments
  • embedded compute
  • communication buses
  • actuator-level control
  • hardware-in-the-loop validation
  • prototype-to-deployment transitions
  • multi-modal robot learning
  • motor-level control optimization
  • efficient model architectures
  • robotics infrastructure
  • high-degree-of-freedom prototype platforms
  • custom actuators
  • precision sensing systems
  • multi-modal robotic foundation models
  • robotics engineers
  • lab teams
  • technical discussions
  • design reviews
  • team leaders
  • fellow scientists
  • compute cluster
  • advanced robotics lab
  • custom actuation systems
  • sophisticated perception systems
  • adaptive manipulation strategies
  • Amazon's massive computational infrastructure
  • rich real-world datasets
  • multimodal perception
  • images
  • videos
  • sensor data
  • diverse real-world scenarios
  • Amazon's global operations

What the JD emphasized

  • original research
  • publishing
  • deploying your innovations into production systems
  • push the boundaries of what's possible
  • take full ownership of turning breakthrough ideas into working systems
  • operate at the intersection of innovative AI research and real-world robotics
  • drive independent research initiatives
  • designing novel frameworks
  • bridge state-of-the-art research with real-world hardware deployment
  • balance innovative technical exploration with hands-on hardware implementation
  • ensure your models and algorithms perform robustly on physical robotic platforms in dynamic real-world environments
  • tackle ambitious problems
  • hardware-aware robotics challenges
  • hardware-in-the-loop validation
  • prototype-to-deployment transitions
  • Design and implement innovative systems and algorithms
  • leveraging our extensive computational and robotics hardware infrastructure to prototype and evaluate at scale
  • solve complex technical challenges spanning motors, actuators, sensors, and learned control
  • Lead technical initiatives from conception to hardware deployment
  • integrate your solutions into physical robotic platforms
  • Transform theoretical insights into practical solutions that perform reliably on real-world robotic hardware
  • building the future of intelligent robotics through ground breaking foundation models and end-to-end learned systems
  • tackle some of the most challenging problems in AI and robotics
  • scale to meet the demands of Amazon's global operations
  • pushing the boundaries of what's possible in robotics
  • seeing your innovations deployed at unprecedented scale
  • Experience building machine learning models or developing algorithms for business application and applying theoretical models in an applied environment

Other signals

  • foundation models
  • robotic intelligence
  • hardware-aware robotics