Hacker News new | ask | show | jobs
by keithwinstein 1626 days ago
Couple of suggestions:

- A lot of people have written really good advice about the application process; see https://blog.keithw.org/2016/02/stock-advice-for-undergradua... for a compilation.

- In the U.S., university faculty usually have substantial influence (and, depending on the department, sometimes close to "they 100% make the decision yes or no") on the PhD admissions decisions among applicants to their department's PhD program who name the faculty member in their application materials. So if you want to come and do a PhD as a "non-traditional" applicant, your real task is to have some faculty members decide they want to hire you as their student.

- If you want to start doing research as a non-traditional applicant, the faculty member's main questions are likely to be things like, "What can this person do for me and my research group? Will this person be a net asset and a contributor of ideas and perspective and get-it-done-ness?", but...

- The prospect of a "mature" PhD applicant can be tantalizing to a faculty member because (a) it's helpful to have people in a research group who have a different perspective and new ideas (about problems of interest, and ways to solve them), (b) older applicants might be assumed to be more mature and have better self-knowledge about what kind of environment they need around themselves to be happy/productive, and (c) applicants from the industry often have skills that are useful to a research group but not necessarily widespread in academia.

- But... the PI may also wonder, "Is this person going to enjoy and survive in an unstructured research environment where they are my apprentice, but essentially their own boss most of the week, and they have to sustain their own motivation in a multi-year research project where success is uncertain and goals often shift?" or perhaps, "If this person was living on $300,000 a year in the industry, do they really want to come here and work for me as a doctoral student for $62,000 for 5-6 years, or are they going to leave and go back to industry when the going gets tough in year 3? Are they just coming here to meet people to form a startup with and I'll have wasted a year or two training them?"

- Some of the ways to de-risk that person's hiring decision are to (a) get your foot in the door in research somehow -- this may require volunteering for free in a research group if you are already out of school, (b) demonstrate promise in research in that group by contributing to a project and ideally a paper submission, and (c) have the PI be able to write you strong letter of recommendation where they write something like, "X was always contributing new ideas and has a great deal of independence/resilience in the face of setbacks/creativity/brilliance/a unique perspective on problems. I would gladly take X as my own student! They remind me of a young [now accomplished researcher] when they were at a similar stage." This kind of letter will be persuasive with the PI's colleagues in the same subdiscipline at other universities, and if the PI really likes you, they will offer you admission as a PhD student at their own institution. Or, you could do something else awesome that demonstrates initiative and creativity and resilience and makes you an interesting applicant -- you don't necessarily have to work in a research lab, but it does make the decision easier.

- For the love of God, please read the faculty member's own website before you email them. Do not send a long email that is all about yourself, your accomplishments, and your interests. Faculty members at well-known universities get tons of those emails. What's interesting to the recipient is usually some question about their work and their interests that provokes thought. If you have a good question about their work, ask it! If you have an idea for how their work might be extended or improved, you might ask if they know of somebody already working on that idea. Faculty members usually love talking about their work and engaging with people who seriously engage with them.

- Other ideas for how to get a sense of what research is like and what you might be interested in (and meet people): attend talks at your local college or university (they are usually open to the public). Go to research conferences in your field of interest and listen to some talks and meet people in the hallways and poster sessions.

1 comments

> If this person was living on $300,000 a year in the industry, do they really want to come here and work for me as a doctoral student for $62,000

Damn $62k as a PhD student? I thought it was more like half that. It wasn’t too long away I was only making a bit more than that in a failed industry career.

I’m thinking about doing one (not in CS though) and this does seem like a helpful blog, though I’ll have to complete undergrad first, which is it’s own beast a non traditional student (I’m convinced the university system doesn’t give a damn about non traditional students outside very specific demographics)

That is a crazy stipend! Maybe somewhere in California + CS + some kind of top-up, but elsewhere, half of that is still relatively generous.

For reference, the NIH stipend for a starting grad student is $26,648; you'll get to $62k after six years...of postdoc.

That's the current salary for a Stanford CS PhD candidate who stays for the summer in a "90%" research assistantship (as is typical in some subdisciplines), but (a) obviously the cost of living here is high, and I think some of that salary has to go back and pay for housing and some portion of health insurance, and (b) many PhD students do an internship in industry during the summer and can end up making more than that.