About the job
About the Institute of Foundation Models
We are an innovative research laboratory focused on the creation, understanding, utilization, and risk management of foundation models. Our mission is to propel research forward, cultivate the future generation of AI innovators, and make significant contributions to a knowledge-centric economy.
Joining our team means engaging in the forefront of advanced foundation model training, collaborating with top-tier researchers, data scientists, and engineers to tackle pivotal challenges in AI development. You will play a crucial role in developing transformative AI solutions capable of revolutionizing industries. Your strategic and creative problem-solving abilities will be essential in positioning MBZUAI as a premier global center for high-performance computing in deep learning, facilitating groundbreaking discoveries that will motivate the next wave of AI pioneers.
The Role
You will be responsible for the comprehensive strategy, design, and execution of our general-purpose world modeling initiatives. Your expertise will be pivotal in translating leading-edge research (such as the PAN framework) into robust, production-ready simulators. You will lead a multidisciplinary team of engineers and scientists while ensuring alignment with IFM’s overarching mission.
Key Responsibilities
Technical Leadership & Vision
- Define and enhance the architecture of the world model utilizing the PAN principles:
1. Multimodal data ingestion
2. Mixed continuous/discrete representations
3. Hierarchical generative modeling supported by a state-of-the-art LLM backbone and diffusion-based predictors
4. Generative loss based on real-world observations
5. Simulation for RL agent training
- Set benchmarks for performance, safety, and evaluation, fostering continuous innovation and improvement.
Project & Team Management
- Manage the planning, resource allocation, and timelines of world model projects.
- Supervise research engineers and scientists (including data curators, RL specialists, and simulator developers) to ensure cohesive progress.
Cross-Functional Collaboration
- Collaborate with agent, reasoning, and deployment teams to integrate world model outputs into various applications (robotics, multi-turn dialogue, autonomous systems).
- Engage with external partners (from academia and industry) to leverage the latest advancements and tools.
Governance & Communication
- Provide updates on project status, risks, and key insights to senior leadership and stakeholders.
- Advocate for best practices in reproducibility, documentation, and transparency.