CI/CD & ML Infra Engineer for Frontier AI Accelerators

OpenAI, Inc.; San Francisco (CA); Sur place; USD 318 000 - 485 000; Plein temps

Avantages offerts par ce poste

Medical, dental, and vision insurance

401(k) retirement plan

Paid parental leave

Flexible PTO

Résumé du poste

OpenAI, Inc. is seeking a Software Engineer for CI/CD & ML Infrastructure in San Francisco.

You will define and build CI/CD systems for frontier models and the software stack on custom AI accelerators, enabling reliable, repeatable production readiness across diverse configurations. The role spans model workloads, accelerator software, and systems infra, with emphasis on scalable pipelines, regression testing, and automation.

Qualifications

• You will define and build CI/CD systems for frontier models and accelerator software.
• You will establish scalable CI/CD, GitOps, and monorepo practices.
• You will design regression and benchmarking pipelines with trustworthy signals.

Responsabilités

• Define the model CI strategy for frontier models running on OpenAI custom silicon.
• Build production CI/CD pipelines for functional regression, performance benchmarking, and release qualification.
• Design test matrices across models, configurations, hardware generations, and deployment environments.
• Create baselines, regression detection, and triage workflows with clear ownership for failures.
• Establish GitOps practices for reproducible configuration and environment consistency.
• Shape monorepo architecture and developer workflows at scale.
• Develop scalable orchestration, artifact management, and observability for accelerator-backed CI.
• Collaborate with model, compiler, kernel, runtime, firmware, validation, and hardware teams to automate gates.

Connaissances

CI/CD systems

Python

Go

Rust

C++

Description du poste

OpenAI, Inc. is seeking a Software Engineer for CI/CD & ML Infrastructure in San Francisco.

You will define and build CI/CD systems for frontier models and the software stack on custom AI accelerators, enabling reliable, repeatable production readiness across diverse configurations. The role spans model workloads, accelerator software, and systems infra, with emphasis on scalable pipelines, regression testing, and automation.