Staff Engineer - Multimodal World Models & AI Research

Veeda AI; Seattle (WA); On-site; USD 140,000 - 220,000; Full time

Job summary

Veeda AI is building the next generation of multimodal foundation world models for Physical AI. We are a small, fast-moving team designing and evaluating generative and predictive models across visual and spatial modalities.

The role focuses on end-to-end training, rigorous experimentation, and collaboration with engineers to translate research into reliable capabilities for embodied intelligence.

Qualifications

• PhD or equivalent experience in a technical field.
• Experience training foundation models end to end with strong PyTorch skills and distributed training.
• Ability to independently design, execute, and analyze ML experiments from hypothesis to conclusion.
• Experience developing evaluations that reveal model limitations and inform research decisions.

Responsibilities

• Design, train, and scale generative and predictive models across visual and spatial modalities.
• Develop methods for learning from diverse data sources and conditioning signals into model behavior.
• Improve consistency, robustness, and generalization in models of complex temporal and spatial structure.
• Explore post-training and model optimization techniques to improve quality and efficiency.
• Develop rigorous evaluations and conduct controlled experiments to guide modeling decisions.
• Collaborate with data, systems, and robotics engineers to translate research ideas into capabilities.

Skills

PhD in CS/Engineering

End-to-end ML model training

PyTorch

Experiment design & ablation

Model evaluation development

Education

PhD in Computer Science, Engineering, or related field

Tools

PyTorch

Distributed training frameworks

Job description

Veeda AI is building the next generation of multimodal foundation world models for Physical AI. We are a small, fast-moving team designing and evaluating generative and predictive models across visual and spatial modalities.

The role focuses on end-to-end training, rigorous experimentation, and collaboration with engineers to translate research into reliable capabilities for embodied intelligence.