Job Hunt as a PhD in AI

Some notes from Nathan Lambert on job hunting with a Ph.D. from Berkeley.

Job Process Reflections

  • High yield: Candidates from prestigious schools can experience high response rates to job applications, allowing more time for reflective interview preparation rather than application quantity.
  • Waiting game: Rejections are usually received before acceptances, suggesting a pattern of companies responding faster to non-preferred candidates; it's common and provides an opportunity to network and hone interview skills.
  • Networking value: The AI research community is small and offers the chance to meet high-quality contacts, adding long-term value beyond immediate job prospects.
  • Research scientist titles: These carry prestige and can affect career paths significantly, with companies like OpenAI offering more neutral titles like "Member of Technical Staff" that can have varied recruitment pipelines.
  • Postdoc consideration: A postdoc is suitable for some, opening doors especially in academia; however, for those with industrial aspirations, it's often better to proceed without unless a postdoc offers a uniquely supportive environment for growth.

Interview Preparation

By roles

  • The paper-writing research scientist interviews: all about evaluating your vision, your place in the broader research community, and ability to work with the team.
  • The product-focused research scientist interviews: mix between faculty search self-promotion and practical engineering questions, in reality they will do the same work as other engineers.
  • The machine learning engineer interviews: pushed more on practical problems. More machine learning fundamentals interviews (things like logistic regression, probability fundamentals, vector calculus, and math puzzles).
  • The robotics engineer interviews: fun systems questions. extremely specific to robotics, less time talking about ML or RL.

Most common questions

  • What do you want to work on when you show up? Who do you want to work with? If you had unlimited compute what would you work on?
  • What was a past project you worked on? What was the impact? What did you learn?
  • What styles of communication do you prefer? Do you work in teams? Do you value mentorship?

Actionable Tips

  • Coding Practice: Consistent coding practice, such as using LeetCode premium.
  • ML Background Interviews: While tricky, preparing for "ML Background" interviews (math tricks and basic ML tradeoffs) through coursework study can be advantageous.
  • Job Talk Preparation: Emphasize a strong narrative and vision in your job talk, focusing less on technical detail and more on engaging storytelling.
  • Research Agenda: Clearly outlining a research agenda with timelines and goals can clarify your process and intentions for starting new projects.
  • Saying No: Be willing to decline unreasonable demands from companies; most are flexible and aim for a positive candidate experience.
  • Interview Topics: Inquire about interview preparation topics; recruiters might provide specific areas to focus on, such as programming concepts or problem-solving domains.