Software Engineer, GPU Inference
Description
The Sora team is pioneering multimodal capabilities for OpenAI’s foundation models. We’re a hybrid research and product team focused on integrating multimodal functionalities into our AI products, ensuring they are reliable, user-friendly, and aligned with our mission of broad societal benefit.
About the Role
We’re looking for a GPU Inference Engineer to contribute to improvements in model serving efficiency for Sora. This is a high-impact role where you’ll drive initiatives to optimize inference performance and scalability. You’ll also be engaged in model design, to help assist our researchers in developing inference-friendly models.
This role is critical to scaling the team’s broader goals - it will directly enable leadership to focus on higher-leverage initiatives by building a stronger technical foundation.
In this role you will:
Perform engineering efforts focused on improving model serving, inference performance, and system efficiency
Drive optimizations from a kernel and data movement perspective to improve system throughput and reliability
Partner closely with research and product teams to ensure our models perform effectively at scale
Design, build, and improve critical serving infrastructure to support Sora’s growth and reliability needs
Qualifications
Have deep expertise in model performance optimization, particularly at the inference layer
Have a strong background in kernel-level systems, data movement, and low-level performance tuning
Are excited about scaling high-performing AI systems that serve real-world, multimodal workloads
Can navigate ambiguity, set technical direction, and drive complex initiatives to completion
This role is based in San Francisco, CA. We use a hybrid work model of 3 days in the office per week and offer relocation assistance to new employees.