Senior Software Engineer, Computer Vision (Knowledge Distillation)

Lead the design and implementation of high-efficiency engineering systems for generative AI tasks, focusing on training, optimization, and deployment with hardware-efficient practices.
Develop acceleration technologies for text-to-image and text-to-video models through knowledge distillation, architecture redesign (dynamic MoE routing, sparse attention), and low-bit quantization.
Advance diffusion model innovation by reducing sampling steps and performing latent optimization to boost generation quality and speed.
Tackle problems in image and video generation, editing, and understanding, delivering scalable, production-ready features.
Own code quality, testing, performance benchmarking, and cross-team collaboration to ship end-to-end solutions.
Require strong software engineering fundamentals in C++/Python, deep experience with PyTorch/DeepSpeed, and a track record in fast-paced environments.