Staff Scientist: Diffusion World Models for Robotics

Odyssey; Zürich; Vor Ort; CHF 120.000 - 180.000; Vollzeit

Zusammenfassung

Odyssey is an AI lab pioneering general world models for robotics and beyond. We seek researchers and engineers with deep ML curiosity who can push diffusion-based world models toward real hardware deployments.

You will work at the intersection of diffusion, world models, and model-based RL, building pipelines, experiments, and policy integrations for cutting-edge robotic systems. Zurich in person work is common and collaboration is global.

Qualifikationen

• PhD or equivalent research experience is required or strongly preferred.
• Hands-on experience with diffusion/world models or model-based RL is essential.
• Proven ability to own projects end-to-end and work across ML pipelines.

Aufgaben

• Learn how interactive world models integrate data, diffusion backbones, action conditioning and downstream policies.
• Implement state-of-the-art diffusion and robot-learning algorithms, design losses and RL objectives.
• Run rapid ablations across world models and policies, explore architectures and data mixes.
• Bridge to real robots through data pipelines, fine-tuning and deployment on Robot as a Service platforms.
• Utilize GPUs to improve training and inference efficiency.
• Take ownership of the full ML stack used by Odyssey researchers and product engineers.

Kenntnisse

PyTorch

Diffusion models

World models

Model-based RL

Experiment-driven

Data pipelines

Ausbildung

PhD or equivalent research experience

Tools

Python

TensorFlow/JAX

Jobbeschreibung

Odyssey is an AI lab pioneering general world models for robotics and beyond. We seek researchers and engineers with deep ML curiosity who can push diffusion-based world models toward real hardware deployments.

You will work at the intersection of diffusion, world models, and model-based RL, building pipelines, experiments, and policy integrations for cutting-edge robotic systems. Zurich in person work is common and collaboration is global.