OpenAI

Final interviews are virtual by default with an option to come onsite in San Francisco. The final-round day itself is unusually long, and the loop runs across several weeks of role-specific preparation – coding patterns for engineering tracks, research depth for research tracks – so a candidate is typically running this process alongside ongoing study or work rather than in parallel with many other concurrent loops.

OpenAI publishes three distinct early-career routes on its Emerging Talent page:

• Early-career full-time roles: research, applied engineering, and product positions for candidates with 0–3 years of experience, including recent graduates and self-taught learners.
• OpenAI internships: open to undergraduate and master's students across engineering, business, and other fields, contributing to live projects alongside permanent teams.
• OpenAI Residency: a six-month full-time program for researchers and engineers from quantitative or scientific fields (mathematics, physics, neuroscience) who are not yet specialised in AI, based at the San Francisco HQ with at least three days a week in office and full-time employee pay stated at $18,333 per month plus benefits.

These tracks are not stacked: the Residency is full-time employment that cannot be combined with concurrent enrolment in a degree, while internships are explicitly for current students. A candidate picks the route that matches their stage rather than applying to all three.

The same emerging-talent and residency pages do not publish a separate visa policy for internships or early-career full-time roles, so a non-US candidate applying to those tracks should confirm the per-requisition position with the recruiter at the introductory call rather than assume it. Where sponsorship is offered for the Residency, the role requires physical presence in San Francisco at least three days a week and the firm does not allow Residents to work remotely or from satellite offices, so the visa decision and the relocation decision are effectively one decision.

First, the Residency is a six-month, full-pay, employee-status pathway into a frontier AI lab open to people from adjacent quantitative fields – mathematics, physics, neuroscience – rather than only AI specialists, and Residents who perform well receive a full-time offer at completion. Second, the interview philosophy is explicitly stated as not credential-driven: the same five-stage loop applies to recent graduates and self-taught learners as to senior hires. Third, the Residency requires at least three days a week onsite at the San Francisco HQ and explicitly disallows remote work or work from satellite offices, so a candidate's willingness to relocate to San Francisco is a hard input alongside the salary and full-time conversion path.