OpenAI jobs, judged for you, free

What they do

OpenAI is an AI research and deployment company. Its mission is to make sure that artificial general intelligence benefits all of humanity. Products include ChatGPT, Codex for software engineering, the OpenAI API, and open models. The nonprofit OpenAI Foundation governs the for-profit OpenAI Group.

What they value

• Humanity first.
• Act with humility.
• Feel the AGI.
• Ship joy.
• Find a way, and update quickly.
• Creativity over control, with intense focus.

Where they hire

Most roles are hybrid in San Francisco, three office days a week, with relocation help. Smaller hubs include Seattle, New York, London, Dublin, Tokyo, and Singapore. India roles are in Delhi, Mumbai, and Bangalore, and are almost all business roles. Delhi and Mumbai roles are hybrid, three days a week. A few roles are remote in the US or Canada.

Teams and roles

Engineering spans applied AI, scaling, consumer devices, and research. Backend roles ask for Python, Go, Rust, or TypeScript.

• Applied AI: ChatGPT, Codex, and API products, including monetization, integrity, and observability. Codex work is often in Rust.
• Scaling: training and inference runtimes, kernels, GPU clusters, datacenters, and silicon.
• Infrastructure: Terraform, Kubernetes, Azure, Postgres, and Kafka.
• Devices and hardware: C, C++, or Rust.
• Research engineering: Python and PyTorch.
• Customer-facing applied AI: production Python, evaluations, and systems from prototype to production. India has some of these roles.

Product and design roles are few and mostly in San Francisco. India roles include account directors, partner deployment engineering with AWS, marketing, and strategy and operations.

What they look for

OpenAI is not credential-driven. It looks for experts and for high-potential people who ramp up quickly. Postings ask for ownership from start to finish, comfort with ambiguity, humility, and commitment to safe AI. The steps are a résumé review, calls with a recruiter or hiring manager, and skills tests such as pair coding or take-homes. Final interviews last four to six hours with four to six people, virtual by default. Engineering interviews judge design, code quality, performance, and test coverage. The guide asks candidates to read the OpenAI Charter and recent research.