Google DeepMind exec says 'work like a dog' to land top AI jobs

Vladimir Feinberg, a distinguished engineer at Google DeepMind who leads Gemini pre-training, recently shared tips on how to land jobs at top AI labs like OpenAI and Anthropic. He says competition is fierce, and successful candidates often come from top universities with a strong sense of purpose, math maturity, and perseverance.

Feinberg believes students need to put in extra hours outside class, sacrificing nights and weekends to build skills. There's no shortcut to math maturity, and the best way to get a lab job is to prove you have the specific skills they need.

To solve the chicken-and-egg problem of needing experience to get into a lab, Feinberg suggests looking at the edges of what labs do, like supporting areas for large language models. He also advises being the kind of person colleagues want to see succeed, by finding opportunities where teamwork and complementary skills shine.