Role Overview

The Agent Infrastructure team at OpenAI is at the forefront of AI innovation, creating scalable systems that enable researchers to train agentic models in environments that mimic real-world software engineering workflows. Our custom-built platforms handle extreme compute scales, providing flexible workspaces where AI agents can execute code, debug issues, and develop software autonomously.

In parallel, we maintain OpenAI's core production platform that powers agentic products like Codex, Operator, tool use in ChatGPT, and upcoming agentic offerings. As a Software Engineer on this team, you'll bridge research and product engineering, scaling novel infrastructure from research prototypes to global production systems serving massive user bases.

This hybrid role (3 days/week in office) is based in San Francisco, CA or New York City, NY with full relocation support. You'll tackle the most challenging problems in AI infrastructure: building container orchestration that surpasses Kubernetes, optimizing globally-distributed compute clusters, and creating APIs that serve petabyte-scale training workloads.

OpenAI seeks engineers with proven experience scaling ML infrastructure who thrive on ambiguous, high-impact problems at the intersection of systems engineering, AI research, and product deployment.

Key Responsibilities

Your day-to-day will involve hands-on engineering across the full stack of agent infrastructure:

• Architect and implement custom container orchestration platforms that scale beyond Kubernetes limits for massive AI training clusters
• Design, develop, and maintain high-performance FastAPI and gRPC APIs serving as the core interface for agentic systems
• Leverage Terraform to provision complex, multi-region infrastructure supporting both research experimentation and production workloads
• Partner directly with AI researchers to build and optimize novel training environments for agentic model development
• Push compute clusters to extreme performance limits, identifying and eliminating bottlenecks in distributed training pipelines
• Scale prototype agent capabilities from research demos to production systems handling millions of concurrent users
• Build flexible execution environments that emulate diverse real-world scenarios for agent training and evaluation
• Maintain and evolve OpenAI's production agent platform powering products like ChatGPT tool use and future agentic offerings
• Debug complex issues across virtualization layers, networking stacks, and application runtimes in high-scale environments
• Collaborate cross-functionally with product engineering to integrate agent infrastructure into consumer-facing products
• Optimize system performance across globally-distributed clusters serving hundreds of millions of daily interactions
• Drive infrastructure evolution to support increasingly complex agentic capabilities and training paradigms
• Contribute to platform reliability ensuring 99.99% uptime for mission-critical AI agent services

Qualifications

We're looking for senior engineers with deep expertise in ML infrastructure and systems at scale:

• 5+ years experience building large-scale machine learning training infrastructure
• Demonstrated success scaling systems from prototype to million-scale production deployments
• Expertise optimizing distributed systems performance and eliminating training bottlenecks
• Strong proficiency with Terraform, cloud platforms, and infrastructure-as-code practices
• Hands-on experience with FastAPI, gRPC, and high-throughput API development
• Deep knowledge of container orchestration, virtualization, and compute cluster management
• Proven track record collaborating with researchers on novel AI infrastructure projects
• Exceptional debugging skills across complex, distributed systems stacks
• Strong Python proficiency and systems programming experience
• Passion for solving ambiguous problems at massive technical scale
• Experience with production ML serving systems and agentic/AI deployment platforms

Salary & Benefits

Estimated Total Compensation: $350,000 - $550,000 USD annually (including base salary, equity, and bonuses). Exact compensation depends on experience and location.

Comprehensive Benefits Package:

• Industry-leading health, dental, vision coverage
• 401(k) matching and retirement planning
• Unlimited PTO with flexible vacation policy
• Hybrid work model (3 days/week in SF/NYC offices)
• Full relocation assistance including housing support
• Daily catered meals, snacks, and premium office amenities
• Professional development stipend and conference budget
• Generous parental leave and family benefits
• Wellness programs including mental health support
• Gym memberships and commuter benefits

Why Join OpenAI?

OpenAI is building safe AGI that benefits all of humanity. Our Agent Infrastructure team solves problems no one else has attempted at scales that redefine what's possible in AI. You'll work with brilliant researchers and engineers on systems that power products used by hundreds of millions worldwide.

Unlike traditional tech companies, every engineer at OpenAI has direct impact on our most ambitious projects. Our hybrid model balances collaboration with flexibility, and our SF/NYC offices offer world-class facilities. Join us to shape the future of agentic AI and work on infrastructure that enables the world's most capable models.

How to Apply

Ready to build the infrastructure powering tomorrow's AI agents? Submit your resume and a brief note about your most impactful infrastructure project. We're moving quickly but hiring thoughtfully - top candidates can expect responses within 48 hours.

Locations: San Francisco, CA | New York City, NY
Hybrid: 3 days/week in office
Visa sponsorship: Available for exceptional candidates