🛠 Tech Stack

💼 About This Role

You'll join Lightricks to enhance training throughput for LTX-2, an open-source generative video model. You'll profile and optimize multimodal data pipelines and develop high-performance kernels for distributed systems and AI accelerators. This role pushes the boundaries of AI and machine learning through collaborative engineering.

🎯 What You'll Do

• Profile and optimize training processes for efficiency.
• Develop high-performance TPU/GPU/CPU kernels.
• Collaborate with researchers on model architectures.
• Design and maintain shared codebase with high quality.

📋 Requirements

• Industry experience with ML experiments and multimodal pipelines.
• Strong skills in Python and modern C++.
• Deep understanding of GPU/CPU/TPU architectures.
• Experience optimizing large-scale training workloads.