- Remote, United States
- Remote
- Full-time · Principal
- 10+ years
Lead inference performance for Runware’s AI Factory, owning kernel development, benchmarking, and team leadership across GPU, HPC, and deep learning systems.
Responsibilities
- Own inference performance end to end for the fleet
- Set the model performance roadmap
- Write and tune CUDA kernels, fuse attention and MLP paths
- Build pipelines for new model deployment
- Choose and shape the serving stack
Requirements
- Nice to have: Diffusion and video generation performance
- Nice to have: DiT
- Nice to have: UNet pipelines
- Nice to have: LoRA hot-swapping
- Nice to have: ComfyUI