Ask HN: How to have an academic career after working in industry? | Hacker News Reader

Ask HN: How to have an academic career after working in industry?

Let's say I'm a research engineer at a startup doing mostly machine learning for science & engineering applications. I joined right after my PhD in a natural science. I'm not keen on moving around for postdoc positions so this seemed like a good option.

I am satisfied with the work but I'm an idealist and my long-term dream is still a faculty position at a university.

1. Did I throw a wrench into that by choosing an industry gig over a postdoc? 2. Are there any (recent) examples of research engineers transitioning into a tenured position? 3. What should I optimize for in my current position to better my odds of achieving that?

(I'm EU based & working for a US company in case it matters.)

View on HN

I'm an academic mathematician. The current academic job market is rough overall, and academia is usually not kind to people who have been out of the system for awhile. However, machine learning is hot so you might get lucky.

Interdisciplinary work is also hot -- but it tends to be popular with deans and provosts, and sometimes less so with faculty members on the front lines, who are the ones who actually read through the hundreds of applications.

Keep in mind that academia works differently in different places; in particular, hiring processes are different in the US and the EU.

I love my job, but academia has its downsides. There is a fair amount of admin work, and you'll spend a lot of your time applying for and managing grants. In laboratory disciplines a lot of successful academics are basically managers.

Another downside is that you don't really get to choose where you live. If you're lucky you will get a couple of job offers, probably in random places where you might not have otherwise chosen to live -- whereas in industry I imagine your credentials might be in wide demand.

I would certainly not expect to be hired into a tenured position. If you are very lucky you might be hired into a tenure-track position, but a postdoc seems more realistic to me. You might be able to get back into the system, but I'd be surprised if you could shortcut past the parts you didn't like.

Another possibility is work as an adjunct professor. This is a terrible career, but could be enjoyable as a side gig. Basically, if you'd like to teach some classes on the side, and don't mind lousy pay and no benefits, you might be able to arrange this.

You might also look at work at various "institutes" -- these are not academia, but are perhaps more suitable for an idealist than a startup. Your credentials are likely to be very attractive to them. Here is one example I am familiar with:

https://www.rti.org/

In any case academia varies widely depending on field, country, and type of university. (In the US there is a big distinction between research- and teaching-oriented universities, this might be less of a thing in the EU.) I'd recommend contacting your PhD advisor, your committee, and any other senior academics whom you personally know and asking their advice. You might also browse through this site, which discusses the practicalities of academic careers, and has a pronounced STEM slant:

https://academia.stackexchange.com

Good luck!

Thank you, this is a great answer. You're right, I should contact a senior academic in my area + field for advice. Sometimes you just don't see the obvious solution :) And thanks for the pointer to independent institutes, I didn't have that on my radar.

The plan is to shortcut past the dreaded PostDoc stage, yes. I realize the hubris but I'd rather try and fail than give up entirely.

For example, lecturers (assistant professors) in CS at some UK universities are very sought after. They are even willing to take PhD students at ABD stage.

Salaries are not great compared to industry. Hence, competition is not very high. If you don't aim excessively high in terms of universities, it might be feasible to skip postdoc.

I'd say you have a big advantage over proper postdocs because you have the salary that allows you to build your own 4x3090 training server. Combine that with a multi GPU Jupyter docker image and you'll be able to do experiments over night where most other researchers need to schedule cluster time and wait a few days.

Also, you will most likely have a lot of experience with tools and solutions that aren't known in the research world, so you'll have an edge in out of the box problem solving.

And lastly, it'll be easy for you to come up with research topics because you know what solutions were missing at work. Plus it'll be easy to get industry buy-in for those problems, which is mandatory for some grant money allocations.

Now you just start some interesting research as a hobby and when you have the first results you reach out to people at your old university and publish a paper together with them. And just like that you're an active researcher again :)

BTW I've been doing AI research for fun myself and so far I've "won" Sintel Clean (optical flow) and German speech recognition. So it's doable, don't worry :)

If you want to reach out, my TEVR paper has the email address.

By networking. Meet some professor who hires postdocs and convince them that you will fit their requirements. Learn how transition grants work. Prepare to live a few years with minimum amount of money coming from your job position. Be ready to travel.

(source: I am a software engineer currently working on my PhD after 10 years in the industry, and planning to continue working in the academia afterwards).

Thank you, but this is exactly what I'm trying to avoid. I had several PostDoc offers that I turned down because I'm not interested in moving countries for a temporary position with no immediate prospects. I know this decimates my chances, but there's a part of me that still thinks I might be able to carve out my own way (hence the question).

> 2. Are there any (recent) examples of research engineers transitioning into a tenured position?

Martin Wattenberg and Fernanda Viégas were appointed as professors at Harvard [1], but they had been working in industrial labs that published lots (IBM Research and Google Research), and are also very good at what they do.

> 3. What should I optimize for in my current position to better my odds of achieving that?

When considering who to hire as an academic, the panel will candidates based on factors including their previous research output, ability to teach students, and ability to attract funding (as well as having a clear plan for their research).

Having evidence of these would be useful.

Publications in top-tier journals or conferences would show that you are able to actually do research.

Giving a guest lecture as part of a university course might be useful as evidence that you can teach; the right lecturer might appreciate an offer to give a lecture that provides a more industrial/applied view on topics covered in the course.

Being involved in a successful bid for an R&D grant from your national government or the EU would also be helpful.

[1]: https://www.seas.harvard.edu/news/2020/11/trailblazing-visua...

Thank you, this is tremendously helpful.

If you are open to a broader score of university work then you might consider connecting with the Research Software Engineer Associations in the countries that you would consider working in. They have a lot of networking discussions over Slack. In the US the URL is:

https://us-rse.org/

I have been at a university for 25 years and absolutely love it. I have a staff position as a research software engineer, but also a non-tenure faculty appointment where I teach classes and work with students. It has really been a best of all worlds combination where I love the work and avoid a lot of the stress and committee work.

It is also possible to go somewhere in between: become a professional researcher associated with a lab.

Upsides:

- Your technical skills will distinguish you from the crowd

- Working on interesting research problems

- Perks of academic life, and probably a fair amount of independence

- You won't have to write a whole bunch of grants

Downsides:

- Pay is less

- Your job still depends on the lab getting funding (with a few exceptions)

- A lot fewer jobs available overall

- You may not get to choose the problems you work on, like you would if you were a PI (but may get the freedom to choose how you work on them)

Anyway, its the compromise I took, because the standard academic track is a brutal grind, and I have a family to feed.

I have a similar background, I come from the natural sciences, did postdocs and then I moved into the tech work

1) Yes 2) Not that I know about and I know many (referring to biology/natural sciences, not CS/economics/business) 3) Very little. Academia (I am referring to biology and the natural sciences, not science, business and commerce) is as backward as it gets for certain (many!) things. When you get out of academia, you are forgotten. If you try to re-enter academia, you are a nobody.

If anything, having experience from the industry should be a big plus – at least that’s my naive view as someone not in academia.

It isn't. For research faculty the most critical thing is having a large number of well received papers in top venues where you were a major author. This can be achieved in industry but it isn't a default.

Firstly, accept the significant drop in salary.

> I'm EU based & working for a US company in case it matters.

I have heard from a British scientist friend working in the US that quite a lot of positions want US national / permanent resident only - I don't know how true this is in your field, but it's something to check.

Some research organisations like to have people with experience in the industry.

Apply for a job in the academic sphere :)

Reply on news.ycombinator.com