Software Engineer, Identity Infrastructure Engineering

OpenAI OpenAI · AI Frontier · San Francisco, CA · IT

Software Engineer on the Identity Infrastructure Engineering team responsible for creating, deploying, and operating foundational security tools and infrastructure to protect OpenAI's model weights, customer data, and critical systems across multiple cloud environments. The role involves building IAM platform features, multi-cloud security architectures, and automation tooling to support AI research and product development.

What you'd actually do

  1. Build new features for our IAM platform that seamlessly integrate with evolving cloud services, enabling teams to work efficiently while adhering to security best practices.
  2. Drive security innovation by designing tools, processes, and architectures that protect data at scale and reinforce a secure development culture across the organization.
  3. Collaborate cross-functionally with researchers, engineers, and compliance teams to address security requirements for multi-cloud deployments, large-scale model training, and emerging AI use cases.
  4. Implement and refine access policies that strike the right balance between enabling rapid experimentation and protecting high-value assets, including model weights and customer data.
  5. Troubleshoot complex identity or access issues across distributed systems, ensuring minimal downtime and a safe environment for AI research and product teams.

Skills

Required

  • Proficiency in programming languages such as Python, Go, or similar
  • Experience with modern cloud infrastructure (AWS, Azure, GCP)
  • Familiarity with industry-standard security protocols (OAuth, SAML, OpenID Connect)
  • Knowledge of threat modeling, risk assessment
  • Ability to embed security features throughout the software development lifecycle

Nice to have

  • Background in building secure systems—from core IAM services to orchestration layers that manage credentials, roles, or policies at scale
  • Excellent collaboration skills

What the JD emphasized

  • protect our model weights, customer data, and critical systems
  • multi-cloud deployments
  • large-scale model training