SpAItial is pioneering the next generation of World Models, pushing the boundaries of generative AI, computer vision, and the simulation of reality.
We are moving beyond 2D pixels to build models that natively understand the physics and geometry of our world.
Our mission is to redefine how industries, from robotics and AR/VR to gaming and cinema, generate and interact with physically-grounded 3D environments.
We're seeking a Research Scientist to train the policies that turn a world model into a robot that acts.
You will own vision-language-action (VLA) and world-action models (WAM) end to end, starting, including data, backbone, action representation, training runs, and the evaluation that tells us whether a policy is genuinely competent or merely lucky.
A world model that understands geometry and physics still doesn't act on its own; the policy is what closes that gap.
This is a senior, hands-on research role for someone who has already trained manipulation policies that worked, and who can say precisely why the ones that didn't failed.
ResponsibilitiesOwn the training pipeline for vision-language-action (VLA) and world-action models (WAM) end to end, from data to a policy running on a robot.
Contribute to setting the technical direction for embodied research at SpAItial.
Close the sim-to-real gap through domain randomization, system identification, and calibration, and build evaluation that predicts real-world transfer.
Adapt VLM backbones for control: encoder choice and adapter strategies, co-training.
Curate and weight the training mix across heterogeneous robot datasets, spanning differing embodiments, action spaces, and sensor setups.
Design action representation and decoding, including tokenization, chunking, diffusion, and flow-matching action experts.
Build the world-model components that predict future observations conditioned on action.
Run post-training: supervised fine-tuning onto target embodiments, and RL for robustness beyond demonstrations.
Key Qualificat.