The bet was simple but audacious: take brilliant people from adjacent fields like theoretical physics and bioengineering, embed them with research teams for six months, and see if they can push AI forward. According to OpenAI, that bet just paid off handsomely.
The lab announced that its inaugural cohort of Fellows has wrapped, with all six participants authoring or contributing to papers by the end of their apprenticeship. Each Fellow started as a machine learning beginner. By the finish line, they were core contributors. It’s a talent pipeline experiment that challenges the conventional wisdom about who gets to do meaningful AI research — and how quickly they can start.
Christine Payne, a Fellow with a background in physics and neuroscience, put it bluntly: “The program gave me the chance to dive into AI research full-time, surrounded by people who pushed me to think rigorously about hard problems.” That intensity seems to be the secret sauce. Rather than a traditional classroom track, Fellows were thrown into the deep end alongside existing OpenAI teams, investigating novel research ideas from day one.
OpenAI isn’t keeping the curriculum locked up either. The organization open-sourced part of the introductory material as “Spinning Up in Deep RL” — a resource packed with clean code examples, documentation, and exercises designed to turn a motivated newcomer into a skilled reinforcement learning practitioner. The next cohort is already in progress, and several of these summer Fellows are now full-time technical staff. It suggests the program doubles as an extended, high-stakes job interview. Applications for the winter class are closed, but another call is coming later in 2019.
