Google DeepMind’s senior engineer Vladimir Feinberg has advice for college students looking to work at frontier artificial intelligence (AI) labs such as OpenAI, Anthropic, and Google DeepMind. Sharing his guidance in a blog post titled “How to Land a Frontier Lab Job”, he wrote that success in the field requires intense dedication, rigorous academic training and sustained effort.Explaining why it’s so tough to succeed in this field, Feinberg, an engineer who leads Gemini pre-training at Google DeepMind, noted that competition for positions at leading AI research labs has intensified as some of the most academically accomplished students pursue careers in artificial intelligence. “Work like a dog. Take difficult, proof-based classes. Code, obviously. Use AI for what you already know how to do, only, but aggressively so,” Feinberg wrote while outlining what he would tell himself if he were entering college today.In a recent episode of The Peterman Pod, Feinberg said that after publishing his blog post, he heard from employees at both OpenAI and Anthropic who agreed with the advice, suggesting that the recommendations resonate across multiple frontier AI labs.
Google DeepMind senior engineer’s advice for college students looking to work at OpenAI, Anthropic and other AI labs
For students entering college and hoping to eventually work at leading AI labs, Feinberg advised fully committing to academic and technical development.Giving reference from his own experience, Feinberg wrote, “Give up your weekends and nights. Burned into my mind is the typical workflow my college friends and I have gone through. We would start from the very morning of Saturday with two big quadruple-shot Panera iced coffees until late, then come back and do it again the following day, hoping to finish early enough to trudge back to our rooms on the other side of campus to get to sleep on time for the start of another week of psets. Rolling deadlines collapse on each other with nothing but another such weekend to look forward to. Outside, clear blue skies host a warm sun shining over the Cottage Club, an eating club we never set foot in, where, locked in our tower atop Lewis Library, the blasting music from its backyard reverberated off the windows as Sunday Funday proceeded differently for those inside versus outside. Not for you. Get that soft sponge in your head, which was mostly a primitive state machine tuned by evolution to hunt and eat and fuck, to think abstractly. Learn to build thought.”
Why Google DeepMind senior engineer thinks it’s difficult to get into top AI labs
According to Feinberg, securing a role at a frontier AI lab is difficult because candidates compete against highly accomplished undergraduate and PhD students who are already publishing machine learning research, participating in mathematics and programming competitions, and building professional networks at major AI companies.He noted that many of the students now targeting companies such as OpenAI, Anthropic and Google DeepMind previously might have pursued careers at quantitative finance firms such as Citadel and Jane Street.Feinberg said there is a group of students who gain an early advantage by understanding which problems matter in AI research and spending years sharpening their technical skills.“There is always a vanguard of elite college students (both undergrads and PhDs) who do ML research in top-tier conferences, math and programming competitions, and already have connections to these labs through older classmates or friends,” he wrote.According to Feinberg, these candidates often stand out for three qualities: intent, mathematical maturity, and grit. He argued that the ability to solve complex problems, focus on important research areas, and handle demanding workloads is a strong indicator of future success.
