Research Scientist - World Model

Luma
Redwood City
Workplace: HybridFull timeFunction: Data Science & Machine LearningEducation: phdSkills: ["Research","Publication","Open-source contribution"]

Invent next-generation world-model architectures and controllability mechanisms that make generative video models interactive, physically faithful, and useful for embodied reasoning. Own success metrics such as physical fidelity, long-horizon coherence, and action-following, and run scaling studies to identify compute/data/architecture tradeoffs. Publish frontier research and contribute to open-source releases that serve as a long-term deliverable.

Loading

Loading job details...

Preparing the role view and application actions.

FursaFursa
Luma
Luma
2 weeks ago

Research Scientist - World Model

✓ Verified Job

Canonical indexed version, validated from employer's careers page.

Source: Company careers pageValidated by: Fursa AI
Last checked: 6 hours agoStatus: Live

Job Summary

Invent next-generation world-model architectures and controllability mechanisms that make generative video models interactive, physically faithful, and useful for embodied reasoning. Own success metrics such as physical fidelity, long-horizon coherence, and action-following, and run scaling studies to identify compute/data/architecture tradeoffs. Publish frontier research and contribute to open-source releases that serve as a long-term deliverable.
Location: Redwood City
Workplace: Hybrid
Employment Type: Full time
Job Function: Data Science & Machine Learning

Key Responsibilities

  • •Invent next-generation world-model architectures focused on controllability and physical consistency (diffusion, transformer, autoregressive, or hybrid).
  • •Develop controllability mechanisms such as action conditioning, view conditioning, and long-horizon rollouts for agent stepping.
  • •Define and own metrics including physical fidelity, long-horizon coherence, action-following, and downstream usefulness for policy training.
  • •Run scaling studies to identify where compute, data, and architecture improve outcomes.
  • •Publish research at the frontier and contribute to open-source releases as a long-term deliverable.

Key Requirements

  • •PhD or equivalent research record in ML, computer vision, robotics, or a related field.
  • •Deep expertise in at least one of: large-scale generative modeling (video/3D/world), self-supervised representation learning, or model-based RL.
  • •Strong PyTorch and large-scale training experience on multi-node clusters.
  • •A research record the field recognizes (top-venue publications and/or widely used open releases).
  • •Experience relevant to world models, model-based RL, generative video, neural simulation, or 4D scene representations (nice to have).
Experience:Computer visionRoboticsGenerative videoWorld modelsModel-based RLOpen source
Education:PhD / Doctorate
Skills:ResearchPublicationOpen-source contribution
Tech Stack:PyTorchDiffusionTransformerAutoregressiveGenerative modelingSelf-supervised learningModel-based RLLarge-scale training

Company Brief

Luma
Develops AI-powered tools for capturing, editing, and rendering high-quality 3D scenes from photos and videos, enabling creators to generate photorealistic 3D assets and spatial experiences.
Industry: AR/VR & Spatial Computing
Website