Research Engineer, World Models

What you'll do

• Design and scale generative and predictive models for driving scenes and controllable simulation.
• Translate research prototypes into distributed training and inference pipelines; optimize efficiency, speed, and reliability.
• Build large-scale training data pipelines and maintain the quality and stability of model infrastructure.

What they ask for

• Strong Python and PyTorch or JAX skills, software-engineering fundamentals, and extensive experience with distributed training and large-scale model deployment.
• Master's degree in Computer Vision, Machine Learning, Robotics, or a related field, or equivalent industry experience in model development and scaling.
• Experience building and deploying generative or predictive physical-world models for real-world applications, with attention to scale, efficiency, and robustness.
• Experience collaborating in a close-knit research and engineering team, with strong communication skills.

Nice to have

• Experience with large-scale ML training infrastructure such as cloud platforms, Kubeflow, or Ray; or efficient serving and deployment tools such as ONNX or TensorRT.
• Publications or research at a top ML/CV/Robotics conference such as CVPR, ECCV, or NeurIPS.

Pay

$155,000 - $269,000 USD (base salary)

Tools

Python, PyTorch, JAX, Kubeflow, Ray, ONNX, TensorRT

AI focus

Develop video, multimodal, generative, and predictive world models for temporal reasoning, scene understanding, traffic prediction, controllable simulation, and autonomous-driving planning, testing, and training.