Machine Learning Engineer - World Modeling

About the role

The Institute of Foundation Models (IFM) at MBZUAI is a research lab focused on building foundation models from scratch and understanding them deeply. This Machine Learning Engineer position joins the PAN world model project, which builds foundation models that simulate, predict and interact with the physical world. You will own the engineering backbone of PAN: large-scale video and simulation data pipelines, distributed training for diffusion transformers, game-engine simulation environments and world-model integration into robotics.

Key responsibilities

• Build and maintain large-scale video and simulation data pipelines, including collection, cleaning, annotation and filtering.
• Develop and optimize distributed training systems for large-scale diffusion transformers and world models.
• Build interactive simulation environments (e.g. Unreal Engine, Blueprint-based gyms, game integrations) for training and evaluation.
• Integrate world action models and video action models into robotics systems.
• Optimize inference and serving for real-time interaction and turn research prototypes into reliable, reproducible systems.

Required profile

• BSc or above in Machine Learning, Computer Science, Robotics or a related field, or equivalent industry experience.
• Hands-on experience with state-of-the-art video generative models and world models (e.g. Cosmos-3, LTX 2.3, Self-Forcing, Lingbot-World).
• Deep expertise in at least one of: full-stack data pipelines, model training and infrastructure, rendering engines and simulation, or world action models and robotics.

Required skills

• PyTorch and debugging of training and inference failures (memory issues, deadlocks, I/O bottlenecks).
• Proficiency with AI coding agents and web-based coding tools (e.g. Claude Code, Codex, Cursor).
• Nice to have: diffusion inference acceleration, web scraping/crawling (scrapy, playwright, selenium), Unreal Engine C++/Blueprint development, and deploying world models in robotics or embodied-AI settings.