Principal Reality Model Research Engineer

Movendi Inc.; San Francisco (CA); Hybrid; USD 170,000 - 210,000; Full time

Job summary

Movendi Inc. seeks a Principal Reality Model Research Engineer to push beyond transformer-centric models, designing latent representations of physical reality and building temporal prediction mechanisms.

You will explore hierarchical spatial reasoning and publish internal research notes, contributing to cutting-edge AI research in a fast-paced environment. The ideal candidate holds a PhD or exceptional equivalent, is a deep learning researcher with strong mathematics, and is proficient in

Qualifications

• PhD or exceptional equivalent required.
• }, {

Responsibilities

• Invent new neural architectures beyond today's transformer-centric approaches.
• Design latent representations of physical reality.
• Build temporal prediction mechanisms.
• Develop hierarchical spatial reasoning.
• Publish internal research notes.

Skills

Deep learning research

Strong mathematics

PyTorch

Transformer knowledge

World models

Diffusion models

Reinforcement learning

Latent variable models

Physics curiosity

Education

PhD or exceptional equivalent

Tools

CUDA

Job description

Principal Reality Model Research Engineer

• Invent new neural architectures beyond today's transformer-centric approaches.
• Design latent representations of physical reality.
• Build temporal prediction mechanisms.
• Develop hierarchical spatial reasoning.
• Publish internal research notes.

Ideal profile

• PhD or exceptional equivalent.
• Deep learning researcher with strong mathematics.
• PyTorch. CUDA is a plus.
• Familiar with world models, diffusion, transformers, reinforcement learning, and latent variable models.
• Curious about physics rather than only language.