Research Scientist - World-action Foundation Model, Robotics

Applied Intuition Applied Intuition · Robotics · Sunnyvale, CA · AI Research

Research Scientist focused on pretraining world-action foundation models for robotics and autonomous systems, incorporating various modalities like vision and physics. The role involves research publications and collaboration with engineering teams for deployment.

What you'd actually do

  1. Conduct research on pretraining world-action foundation model with various world modalities including vision and physics associated with ego actions, serving the purpose for both robot action and simulation world generation.
  2. Diving into relevant topics such as vision-physics modality association, feed-forward Gaussian splatting, world foundation model, human data incorporation, language modality, and spatial reasoning, deformable object modeling and simulation.
  3. related topics including 3D/world-action foundation model, multi-modal pretraining, feed-forward Gaussian splatting, world foundation model with applications to autonomous driving, and fundamental topics on 3D vision and generation with broader applications
  4. Work closely with other Research Scientists and interns on research publications for submission to top-tier conferences
  5. Collaborate with Research Engineers and engineering teams to test and deploy algorithms to our autonomy and simulation products

Skills

Required

  • Python
  • Pytorch
  • computer vision
  • robotics systems
  • distributed machine learning model training

Nice to have

  • 3D foundation model and pretraining
  • Multi-modal foundation model
  • Feed-forward Gaussian splatting and reconstruction
  • World foundation model and generation
  • 3D/multi-view end-to-end models for autonomous driving or robotics
  • Human data processing and incorporation

What the JD emphasized

  • publications in top-tier conferences or journals in the fields of computer vision, machine learning, and robotics

Other signals

  • foundation model
  • robotics
  • pretraining
  • vision-physics modality association