🛠 Tech Stack

💼 About This Role

You'll define model quality for LTX-2, an open-source generative video model. You will design automated judges and reward models that guide training and benchmarks. This role blends research rigor with product intuition to ensure world-class video generation.

🎯 What You'll Do

• Evaluate model checkpoints to guide training experiments and architectural decisions.
• Develop and run rigorous benchmarks for release candidates.
• Build automatic metrics and reward models for complex video attributes.
• Perform root-cause analysis on model failures to enable targeted fixes.
• Collaborate with platform engineers to scale evaluation on GPU clusters.

📋 Requirements

• Master’s or PhD in Computer Vision, ML, or related field.
• Strong software engineering skills and comfort in ML training environments.
• Deep expertise in evaluation methodology and statistical rigor.
• Sharp perceptual intuition for video and audio quality.