Applied Scientist Iii, Rbks AI

Amazon Amazon · Big Tech · CA, ON +1 · Applied Science

The RBKS AI team at Amazon is seeking Applied Scientists to innovate AI features for Ring and Blink cameras, focusing on the intersection of computer vision, generative AI, and ambient intelligence. The role involves productizing research into advanced computer vision and multimodal GenAI models for video understanding, object detection, and real-time applications, with an emphasis on privacy-preserving, efficient fine-tuning, and on-device/in-cloud inference. The goal is to ship AI solutions that enhance home security for millions of customers.

What you'd actually do

  1. Design and develop advanced computer vision and GenAI models and algorithms for comprehensive video understanding, including but not limited to object detection, recognition and spatial understanding
  2. Develop privacy-preserving CV and GenAI models and systems, focusing on efficient fine-tuning and on-device and in-cloud inference
  3. Map product requirements into science solutions and deliver high-quality science artifacts that ship to products
  4. Collaborate with scientists, engineers, product/program managers and other cross-functional teams
  5. Provide technical leadership on AI products/features, and develop and mentor junior scientists on the team.

Skills

Required

  • PhD, or Master's degree and 8+ years of applied research experience
  • developing and optimizing computer vision models
  • multimodal LLMs
  • Python
  • computer vision frameworks
  • GenAI frameworks

Nice to have

  • dealing well with ambiguity
  • prioritizing needs
  • delivering measurable results in an agile environment
  • hardware-software co-design for CV and GenAI applications
  • visual transformers
  • diffusion models
  • multimodal generation
  • real-time computer vision systems
  • optimization techniques
  • efficient training and deployment of vision models and multimodal large language models
  • Published research in top-tier conferences (CVPR, ICCV, NeurIPS, ICML) focusing on computer vision and/or GenAI

What the JD emphasized

  • advanced computer vision
  • multimodal GenAI
  • video understanding
  • object detection
  • recognition
  • spatial understanding
  • privacy-preserving
  • efficient fine-tuning
  • on-device inference
  • in-cloud inference
  • real-time computer vision systems
  • efficient training
  • deployment of vision models
  • multimodal large language models

Other signals

  • productizing research
  • shipping AI features
  • multimodal GenAI
  • computer vision