Gen AI Staff Researcher - Generative World Models

About the role

GM Israel (Herzliya) is hiring a Gen AI Staff Researcher for its Foundational AI Research team, part of the GM AV AI Research organization. The team develops next-generation autonomous driving, L3 and beyond, and the role contributes to generative world models focused on diffusion models for autonomous driving. The researcher designs and trains models that learn driving-environment dynamics and generate realistic, temporally consistent future scenarios conditioned on sensor observations, actions and scene context.

Key responsibilities

• Contribute original research and technical expertise to generative world models, focusing on diffusion-based modeling of driving environments.
• Design, build, train and evaluate large-scale diffusion models for video and multimodal generation, emphasizing temporal consistency, physical plausibility and controllability.
• Implement and optimize large-scale training on distributed, multi-GPU infrastructure, improving stability, throughput, memory efficiency and scalability.
• Develop action- and context-conditioned world models that generate diverse future scenarios for counterfactual exploration, closed-loop simulation and long-tail scenario coverage.
• Develop and evaluate data and training strategies for fleet-scale datasets, including curation, pretraining, fine-tuning and targeted generation of challenging scenarios.
• Develop evaluation methods linking generative quality and scenario coverage to downstream autonomy KPIs.
• Collaborate with engineering teams to turn research into scalable simulation and synthetic-data solutions, publish at top-tier conferences and journals, and incubate technologies for L3 autonomy.

Required profile

• PhD in Computer Science, Electrical Engineering, Robotics or a related field; excellent M.Sc. graduates will be considered.
• Over 3 years of research experience in generative AI, computer vision, machine learning or related areas.
• Strong publication record at top-tier AI/ML conferences and journals.

Required skills

• Deep expertise in diffusion models and hands-on experience with image, video and multimodal generative models.
• Hands-on experience with large-scale distributed training of deep learning models.
• Understanding of world modeling and spatiotemporal learning.
• Excellent Python coding and proficiency with modern AI frameworks such as PyTorch.