One odd job I do is help some companies by reviewing their AI job reqs. One of the common mistakes I notice HR/hiring managers make in hiring deep learning talent is to put “PhD preferred” in product AI research job req. This goes wrong in so many ways:

The best DL talent (for most companies) don’t need a PhD to do the right things for your company’s product goals. Adding “PhD preferred” makes many otherwise qualified folks not even consider.

You might think, “preferred” implies optionality, but that’s not what the candidate may see. They may lack self-esteem to interpret things that way, or they may be raised in less individualist cultures, where following “rules” is expected.

Plus, by adding “preferred” you are signaling to those otherwise qualified non-PhD candidates that even if they applied and got accepted, they were not the preferred candidate. What a terrible way to onboard! Remember, onboarding begins with the job req.

If I rely on my personal experience, for small startups, PhDs have been a churn risk. A PhD might be okay doing product-focused work at Google/Facebook/etc for a while to earn the FAANG credential, but sorry that’s not true for your startup, esp. in competitive areas like AI/ML.

Also, researchers crave community, esp PhDs. If you already don’t have a couple of exceedingly strong researchers seeding your group and actually have a research mandate, just throwing a “PhD Preferred” will not attract them unless they are super desperate for big$ & fancy titles

And those are usually the worse candidates to hire as money and titles get old and disappear into the background while leaving them exploring why they joined here in the first place. Big mistake.

So why do many founders and hiring managers insist on “PhD Preferred”? 1) faulty logic and pattern matching (“papers we rely on are mostly written by PhD students, so we should hire them”) 2) to make their pitch decks look impressive (“we have 10 PhDs”).

Aside: to expand more on 2), if you are a candidate, run away from such companies. Smells like an acquihire portfolio building to me, esp. if their product/biz/market is not strong or if the tech is not foundational in some way and PhD is essential to acquire skills (e.g. bio).

If you want a PhD on your team, ask why. Be prepared to clearly articulate that. I’m here to argue that even research managers in AI don’t need a PhD, but that’s for another thread.

If you are a PhD, there is so much else you should be looking at before answering to that “PhD Preferred” req, but that’s also a topic for a different thread.

If you are a hiring manager, review your job reqs to see if you are turning away some amazing folks simply for signaling. I don’t have a lot of personal bandwidth for this, but I can help on a case-by-case basis.

Finally, if you are NOT a PhD and thriving in your AI/ML career, share this with other NOT-PhDs and let them know they can apply to a wider range of jobs, including in your companies (tag them).