Sr. Applied Scientist, Aws Just-walk-out Science Team

Amazon Amazon · Big Tech · Seattle, WA · Applied Science

This role focuses on developing novel frameworks and techniques for multi-object tracking, re-identification, person activity understanding, and multi-modal foundation models within the context of Amazon's Just Walk Out technology. The scientist will advance the theory and practice of these areas, create efficient visual processing techniques, and reduce computational/data requirements for visual AI systems. The role requires a strong publication record in top-tier conferences and experience in computer vision, deep learning, and multi-modal foundation models.

What you'd actually do

  1. Develop a novel framework and advance the theory and practice of multi-object tracking, re-identification, person activity understanding, multi-modal foundation model, and generic video understanding
  2. Create innovative techniques for efficient visual processing that can scale to real-world applications
  3. Investigate approaches to reduce the computational and data requirements of visual AI systems
  4. mentor other scientists

Skills

Required

  • PhD, or Master's degree and 6+ years of applied research experience
  • 5+ years of building machine learning models for business application experience
  • Experience with neural deep learning methods and machine learning
  • Track record of publications at top-tier conferences (CVPR, ICCV, ECCV, NeurIPS, ICLR), specifically focusing on multi-object tracking, re-identification, human activity understanding, or multi-modal LLMs
  • Strong background in computer vision, deep learning, and multi-modal foundation mo

Nice to have

  • entrepreneurial
  • wear many hats
  • highly collaborative environment
  • tackle problems that span a variety of domains: computer vision, image recognition, machine learning, real-time and distributed systems
  • comfortable with a degree of ambiguity
  • relish the idea of solving problems that, frankly, haven’t been solved at scale before - anywhere

What the JD emphasized

  • Track record of publications at top-tier conferences (CVPR, ICCV, ECCV, NeurIPS, ICLR), specifically focusing on multi-object tracking, re-identification, human activity understanding, or multi-modal LLMs
  • multi-modal foundation model
  • multi-object tracking
  • re-identification
  • person activity understanding

Other signals

  • Develop a novel framework and advance the theory and practice of multi-object tracking, re-identification, person activity understanding, multi-modal foundation model, and generic video understanding
  • Create innovative techniques for efficient visual processing that can scale to real-world applications
  • Investigate approaches to reduce the computational and data requirements of visual AI systems
  • Track record of publications at top-tier conferences (CVPR, ICCV, ECCV, NeurIPS, ICLR), specifically focusing on multi-object tracking, re-identification, human activity understanding, or multi-modal LLMs