Software Engineer, Productivity - Model Performance
$230k - $385k • Remote • San Francisco • FullTime
Posted 5mo ago
About the job
We are seeking a highly motivated and autonomous developer productivity engineer to join OpenAI's Model Performance organization. This role is crucial for enhancing the efficiency, safety, and confidence of our engineers working on model performance systems. You will contribute to the development of essential infrastructure, including CI/CD systems, testing workflows, and tooling, with a significant opportunity to impact the Triton project and performance-critical engineering work across OpenAI.
Responsibilities
- Improve development workflows for engineers focused on model performance infrastructure.
- Design and enhance CI/CD, release, validation, and testing pipelines.
- Build and maintain tools to boost reliability, iteration speed, and engineering confidence.
- Collaborate with engineers to identify and resolve friction points in testing, debugging, deployment, and development.
- Contribute to infrastructure supporting performance-critical training and inference systems.
- Enhance developer experience within Python-heavy codebases and performance-oriented infrastructure.
- Operate effectively in a high-context, ambiguous environment requiring strong ownership and judgment.
Requirements
- Motivated by enabling others and helping engineers perform at their best.
- Strong experience with CI/CD, developer infrastructure, testing systems, tooling, or build/release workflows.
- Highly collaborative, empathetic, and comfortable partnering deeply with technical teams.
- Proficient in Python and experienced in building reliable, scalable developer tools and infrastructure.
- Experience improving large-scale engineering workflows, particularly CI reliability, test infrastructure, and debugging velocity.
- Self-directed and comfortable operating with ambiguity.
- Excited to learn the model performance domain.
- Experience in the PyTorch ecosystem is highly relevant.
- Experience with C++ or Rust is a plus.
- Instinct to fix underlying systems when encountering repeated friction (e.g., slow tests, flaky CI, brittle release processes, painful debugging, unclear validation).
- Pragmatic approach balancing high standards with forward progress.