I've placed a few candidates at OpenAI and coach many more who are mid-process. Here's what I've gathered about the OpenAI recruiter phone screen based on consistent reports.
Format: usually 30 minutes. Phone or Zoom. Done by an in-house recruiter (not a sourcer, an actual recruiter with some technical context).
What they cover:
Background and motivation. They will ask why OpenAI specifically. This is not a formality. Candidates who give generic 'I'm excited about AI' answers get probed harder. Candidates who can speak to specific research, products, or safety work tend to move faster. Doing homework on their recent papers or model releases is useful, not just their main products.
Logistics. Relocation (they do have a SF office preference for many roles), visa status, timeline, compensation expectations. They'll give you a range if you ask; it's worth asking early to save everyone time.
Background sketch. Quick walkthrough of your experience. They're triangulating on level: are you L4, L5, or staff-adjacent? They make that call before the technical rounds start. If you're on the bubble, saying so explicitly ('I know I'm on the border between senior and staff, happy to let the technical rounds calibrate that') is actually fine.
Questions for them. They expect this and it's another signal. Good questions to ask: what does the team this role sits on actually ship, what does success look like in the first 6 months, how much are individual contributors involved in research direction.
One thing I've noticed from candidate feedback: the recruiters are generally well-prepared and responsive. Multiple people mentioned getting follow-up answers to questions the recruiter didn't know off the top of their head. That's rarer than it should be.
Timeline from phone screen to next step: typically 5-10 business days. They run structured scheduling, not ad hoc.