Full-Time
Staff Scientist
World Model
Institute of Foundation Models
Compensation Overview
$300k - $600k/yr
Sunnyvale, CA, USA
In Person
Category
 AI & Machine Learning (2) |
|---|
| , |
Required Skills
Reinforcement Learning
Requirements
- Ph.D. or M.S. with 8+ years in AI research or engineering, specializing in world modeling, simulation, or generative modeling
- Proven track record building large-scale simulators or predictive models for complex environments
- Deep expertise in transformer-based LLMs, diffusion models, and hierarchical latent representations
- Hands-on experience with reinforcement learning frameworks (policy learning, planning with latent dynamics)
- Strong leadership skills: project management, cross-site coordination, and team mentorship
Responsibilities
- Define and evolve the overall world model architecture, drawing on the PAN principles: Multimodal data ingestion Mixed continuous/discrete representations Hierarchical generative modeling with an enhanced LLM backbone and diffusion-based predictors Generative loss grounded in real observations Simulation for RL-based agent training
- Establish performance, safety and evaluation benchmarks, driving continuous improvement
- Oversee planning, resourcing, and timeline for world model projects. Manage research engineers and scientists (e.g., data curators, RL experts, simulator devs) to achieve unified progress
- Partner with agent, reasoning, and deployment teams to integrate world model outputs into downstream applications (robotics, multi-turn dialogue, autonomous systems).Liaise with external collaborators (academia & industry) to incorporate the latest advances and tooling
- Report project status, risks, and key insights to senior leadership and stakeholders.Champion best practices in reproducibility, documentation, and knowledge sharing
Desired Qualifications
- Experience leading multi-location technical teams in fast-paced R&D settings
- Published contributions to world model architectures or simulation benchmarks
- Track record of taking research prototypes into production systems
Company Size
N/A
Company Stage
N/A
Total Funding
N/A
Headquarters
United Arab Emirates
Founded
N/A
Simplify's Take
What believers are saying
- IFM's dedicated teams in Abu Dhabi, Paris, and Silicon Valley drive K2 and JAIS advancements.
- Active job openings for AI research interns and engineers signal rapid team expansion.
- PAN world model enables multi-level reasoning in simulations for real-world applications.
What critics are saying
- OpenAI's o1 surpasses K2 and JAIS by 25% on benchmarks, shifting users in 6-12 months.
- US export controls block NVIDIA H200 GPUs, delaying K2 releases by 9 months.
- Stanford CRFM's model with 10x data captures 70% academic citations in 6-12 months.
What makes Institute of Foundation Models unique
- IFM pioneers open-source K2 Think V2, UAE's sovereign 70B reasoning system released January 2026.
- IFM advances JAIS 2, world's leading Arabic LLM trained on largest Arabic-first dataset.
- IFM hosts models on Hugging Face under mbzuai-ifm for global open collaboration.
Help us improve and share your feedback! Did you find this helpful?
Your Connections
People at Institute of Foundation Models who can refer or advise you
Benefits
Health Insurance
Dental Insurance
Vision Insurance
Paid Vacation
Paid Holidays
Parental Leave
Employee Assistance Program
Life Insurance
Disability Insurance
401(k) Plan
Wellness Program
Flexible Work Hours
Remote Work Options
Hybrid Work Options
Stock Options
Company Equity