Senior Applied Research Engineer - Video

Build and scale production-grade human-centric video foundation models. Own end-to-end applied research and engineering projects: develop latent video diffusion models, design conditioning for pose/emotion/script/camera control, advance distributed training strategies (DDP, FSDP, DeepSpeed, sequence parallelism), improve multi-node training stability, design evaluation frameworks combining automated metrics and human evaluation, optimize inference for low latency/high resolution/cost efficiency, and run controlled ablations. Requirements: strong experience training deep learning models at scale, Python and PyTorch, diffusion models (image required, video preferred), large-scale multi-GPU/multi-node training, distributed training expertise, and experimental design skills.