Job Description

xAI is seeking a multimodal engineer to pioneer AI agents that perceive, reason about, and generate video content, advancing video understanding and generation within agentic systems. The role involves developing planners, designing data, building evaluations, and studying training recipes.

Responsibilities

• Developing agentic planners for short- and long-horizon video generation
• Designing and collecting human/synthetic data
• Developing data generation techniques
• Building evals and reward models for video generation
• Studying training recipes for advancing video understanding/generation and agent training

Requirements

• Track record in improving neural network capability and performance
• Experience in data-driven experiment designs and systematic analysis
• Experience in SFT, RL, evals, and human/synthetic data
• Experience in agentic RL training and video models is a plus
• Strong communication skills

Benefits

• No benefits