So you want to work at OpenAI? Here’s what it takes
Diane Yoon joined OpenAI five years ago, when the startup was a nonprofit organization with a few dozen employees and AI was the stuff of science fiction. She spent a year working as an operations manager, before moving over to human resources. As Yoon grew into her role, OpenAI grew, too: Today, she’s VP of people, overseeing 375 employees. The technology that was a dream at OpenAI’s founding, in 2015, is now in the hands of millions of people.
San Francisco-based OpenAI has had a breakout year, thanks to the runaway success of its chatbot, ChatGPT, which launched in research preview in November and attracted 1 million users in its first week, making it one of the fastest-growing technology products in history. OpenAI’s innovative approach to deploying its technology to mass audiences earned it a place atop Fast Company‘s list of the Most Innovative Companies of 2023. (Read our profile on Mira Murati and how OpenAI innovates.)
OpenAI is partnering with giants like Microsoft to develop generative AI products and going up against the likes of Google, but it remains a relatively small and tightly knit organization. That slow and deliberate approach to headcount growth is by design, says Yoon, giving OpenAI license to slowly and deliberately find the kind of talent that’s eager to do “the most creative work of their lives.”
“Obviously, we have a résumé review process,” she says. “But for a lot of our roles, we explicitly look for experiences that may not land on a traditional résumé.”
For example, OpenAI’s job application often asks candidates to provide an example of “impactful work” that they’ve done in the past. For Yoon, this question is the key to understanding how motivated someone is to push toward new discoveries. “Sometimes folks who have been [hired as] research scientists first implemented one of our models as a side project,” she says. “They didn’t necessarily have experience as a research scientist, but they were fascinated enough with the work to learn [the technology] and implement something on their own.”
OpenAI’s vetting processes are not the most efficient approach to hiring, Yoon acknowledges. “It’s a little bit more white-glove,” she says. “But we found that that kind of investment has paid off for us in a really big way.”
OpenAI bucks hiring norms in other ways as well. “We don’t really care that much about a candidate’s pedigree—the degrees they have, what colleges they attended, what companies they’ve previously worked for,” Yoon says. “We really just care about skill and alignment [to our mission].” At other AI labs, PhDs are the norm. At OpenAI, there are research scientists with bachelor’s degrees, but nothing more. Others joined the organization straight out of high school. (OpenAI CEO Sam Altman is a Stanford dropout, setting the tone.)
Similarly, OpenAI downplays the academic barometers that typically accompany research work. “We’re a research organization at heart, but fundamentally, we’re trying to solve a problem,” Yoon says. “And when you’re trying to solve a problem, which is to make an AGI [artificial general intelligence] that is safe and beneficial for all, it is much more effective to try the simplest things first, like the dumb idea that that might just work and takes very little investment.”
For new hires, OpenAI reinforces its corporate culture through shared daily practices. Values like transparency and collaboration, for instance, are made real in the company cafeteria, where long communal tables set the stage for conversation. “We have this tradition where you go to lunch and sit wherever,” says Yoon. “People from across teams are expected to do a lot of cross-pollination. What’s the paper that I should read? What’s this technique I should share?”
Yoon has watched OpenAI transform over the years—restructuring to become a for-profit corporation with a nonprofit parent, moving to a new headquarters, adding staff. Yet it feels like much the same place as the day she started. “There’s a magic to working here,” she says, “and I think that magic is just you feel that your colleagues are trying to move in the same direction as you.”
(45)
